The Rill Catalog
How to Read This Guide
Slipstream organizes its intelligence into three layers: domains represent strata of reality, from raw geology to system operations. Within each domain, families group rills by thematic purpose—terra covers soil and slope, atmos covers weather and pollen. Rills are the atomic units: each one retrieves, transforms, or analyzes a single slice of data. The sidebar at left mirrors this hierarchy. Click any family name to jump directly to its profile.
Tags appear in three colors throughout. Green tags are system-assigned from a controlled vocabulary of 89 terms describing data domains and source types. Amber tags are compositional archetypes—they mark rills that participate in named flow types like property-intelligence or biodiversity-survey. Purple tags are community-contributed and emerge through use. The real value of this system lives in the composition examples: showing what emerges when rills from different families are stacked together. A soil profile alone is data; combined with flood risk, zoning, and property valuation, it becomes a signal.
Tag System Reference
System Tags — 89 terms, controlled vocabulary
Compositional Archetypes — 14 named flow types
Community Tags — freeform, emergent
Community tags are not pre-defined. They emerge when three or more independent users apply the same term to a rill. Every rill entry includes the slot: populated through use. Proposed tags enter a candidates pool before becoming visible in the catalog.
Terra Firma — The Physical World
Before anything is built, governed, or inhabited, there is the ground itself. Terra Firma is the bedrock layer of the tomographic stack—the physical substrate on which every other signal rests. Soil composition determines what grows. Slope and aspect determine where water flows and where sunlight falls. Seismic history shapes building codes. Light pollution determines whether you can see the Milky Way from your backyard.
These rills pull from USGS, NOAA, EPA, NIFC, and NASA data sources to characterize the raw physical reality of any location. Alone, each measurement is a fact. Stacked together, they become a portrait of place that no single dataset can provide.
terra (earth, land, ground)
What is the ground beneath your feet actually made of? Terra answers the questions that precede every other question about a place: the soil type and drainage, the slope and aspect, the bedrock geology, what the land is used for now and what it was used for before. It reaches into the sky for air quality and light pollution, and into the earth's magnetic field for compass correction. This is the family you query when you need to understand a site before you do anything with it.
Soil composition and drainage, terrain elevation and slope, bedrock geology, land cover classification, property boundaries, wildfire perimeters, seismic event history, light pollution, ambient noise, air quality index, and geomagnetic declination.
USDA SSURGO (soil, free), USGS National Map (elevation, free), USGS SGMC (geology, free), MRLC NLCD (land cover, free), NIFC ArcGIS (wildfire, free), USGS Earthquake Hazards (seismic, free), EPA AirNow (AQI, key required), NOAA NCEI (geomagnetic, free), Regrid (parcels, commercial key).
A soil profile alone is agricultural data. Stack it with h-07-flood-risk and cv-01-zoning-lookup, and it becomes a buildability assessment. Combine t-03-elevation-slope with t-01-helio-study and gr-03-solar-potential, and you have a solar site evaluation that accounts for terrain shading. Layer t-09-light-pollution with co-10-astrophotography-planner and sn-05-natural-soundscape, and you've composed a dark-sky retreat finder. Terra data becomes different things depending on what you stack it with.
Stack t-02-soil-profile + t-03-elevation-slope + t-07-wildfire-history + h-07-flood-risk + cv-01-zoning-lookup for a parcel near Bend, Oregon. The soil says well-drained sandy loam (good for building), but the slope is 18% (grading costs), the parcel sits within a 2019 burn perimeter (insurance implications), and it's in a FEMA Zone X (minimal flood risk). This surfaces a signal that the site has mixed suitability—structurally sound but with fire history that warrants investigation of current fuel loads and defensible space requirements.
| ID | Name | Description |
|---|---|---|
| t-01 | Helio Study | Solar position, daylight duration, and optimal panel tilt angle |
| t-02 | Soil Profile | USDA SSURGO soil type, drainage class, horizon depths, pH, organic matter |
| t-03 | Elevation & Slope | Ground elevation, slope steepness, and aspect direction via USGS |
| t-04 | Geology & Bedrock | Bedrock formation, lithology, and geologic age from USGS SGMC |
| t-05 | Land Use / Land Cover | NLCD land cover classification (forest, developed, wetlands, etc.) |
| t-06 | Property Boundaries | Parcel data, ownership, zoning, and assessed value via Regrid |
| t-07 | Wildfire History | Historical wildfire perimeters from NIFC within a configurable radius |
| t-08 | Seismic Risk | Earthquake history, magnitude, depth, and computed hazard summary |
| t-09 | Light Pollution | Bortle dark-sky class and sky quality estimate |
| t-10 | Noise Profile | Transportation noise estimate from airport and highway proximity |
| t-11 | Air Quality | Real-time AQI, dominant pollutant, and health guidance from EPA AirNow |
| t-12 | Magnetic Declination | Compass variation and geomagnetic field parameters via NOAA NCEI |
Computes solar position, daylight duration, sunrise/sunset times, and optimal fixed-mount panel tilt angle for any location and date. Pure math—no API calls, no rate limits, instant results.
lat, lng (required); date (optional, defaults to today)
Solar azimuth/altitude/zenith at noon, total daylight hours, recommended panel tilt in degrees
Local computation (simplified Solar Position Algorithm). Free, no auth.
Daily. Solar geometry changes slowly—24-hour cache is appropriate.
A homeowner in Portland evaluating rooftop solar runs t-01 for their address in December and June. December shows 8.6 hours of daylight with the sun peaking at 21° altitude—low enough that a neighboring two-story house could cast significant shadow. June shows 15.5 hours at 65° altitude. The optimal tilt is 35°. Combined with t-03-elevation-slope, this surfaces whether the roof pitch already approximates the ideal angle.
t-03-elevation-slope— Slope aspect reveals whether terrain amplifies or blocks solar exposuregr-03-solar-potential— Adds NREL irradiance data to theoretical sun position for actual energy yieldc-07-sun-shadow-map— Projects shadows from buildings and terrain across the sitecu-01-growing-calendar— Light hours shape planting windows for specific cropsco-10-astrophotography-planner— Twilight timing determines dark-sky observation windows
Retrieves detailed soil profile data from the USDA SSURGO database—soil type, drainage class, horizon depths, texture, pH, and organic matter content. The foundation for any serious conversation about what a piece of land can support.
lat, lng (required)
USDA texture class, natural drainage class, array of soil horizons with full properties (depth, texture, pH, organic matter)
USDA Soil Data Access (SSURGO). Free, no auth, no rate limit.
Weekly. Soil surveys update infrequently—7-day cache is conservative.
A prospective buyer evaluating 5 acres outside Eugene, Oregon queries t-02 and finds Jory silty clay loam—well-drained, pH 5.8, high organic matter. This is prime Willamette Valley agricultural soil. Stack with cu-05-soil-requirements to match crops to this specific profile, or with vl-01-property-valuation to understand whether the land is priced as farmland or development potential. The soil doesn't change, but its meaning does depending on the question.
f-02-native-plant-atlas— Matches native species adapted to the specific soil type and pHcu-05-soil-requirements— Maps crop viability to actual soil compositionh-07-flood-risk— Drainage class modulates flood impact—poorly drained soils compound flood riskt-04-geology-bedrock— Subsurface geology explains why the soil is what it isvl-01-property-valuation— Soil quality is a hidden driver of agricultural land value
Retrieves ground elevation from the USGS National Map and computes slope steepness and aspect direction using Horn's algorithm on a local 3×3 elevation grid. The terrain basics that everything else builds on.
lat, lng (required)
Elevation in meters and feet, slope in degrees, cardinal aspect direction (N, SE, W, etc.)
USGS Elevation Point Query Service (EPQS). Free, no auth.
Monthly. Terrain is essentially static—30-day cache.
A trail designer scouting a ridge route in the Columbia River Gorge queries t-03 at waypoints along the proposed path. One segment returns 34° slope facing north. That's steep enough to require switchbacks, and the north aspect means snow lingers longer in spring. Combined with t-02-soil-profile (erosion-prone loess soils on steep slopes) and h-08-precipitation-tracker (heavy winter rainfall), this surfaces a drainage challenge that the trail design needs to address.
t-01-helio-study— Slope and aspect directly modify effective solar exposuret-02-soil-profile— Steep slopes with poor-drainage soils signal erosion riskh-07-flood-risk— Low-elevation areas with minimal slope compound flood hazardki-08-elevation-profile— Elevation along a route reveals cumulative climbing for activity planning
Identifies the underlying bedrock formation and lithology using the USGS State Geologic Map Compilation. The deep history beneath the surface—useful for understanding why the soil, water, and landscape look the way they do.
lat, lng (required)
Formation name, primary lithology (sedimentary, igneous, metamorphic), geologic age (e.g., Cretaceous, Quaternary)
USGS SGMC (State Geologic Map Compilation). Free, no auth. Status: dev.
Monthly. Geologic maps update on geologic timescales—30-day cache.
A well driller in the San Juan Islands queries t-04 and finds Paleocene-age sedimentary sandstone—a relatively good aquifer rock. Combined with h-06-groundwater-level and t-02-soil-profile, this helps estimate drilling depth and likely yield. The geology doesn't tell you how much water you'll get, but it tells you what kind of rock the drill will hit.
t-02-soil-profile— Bedrock parent material explains soil characteristicsh-06-groundwater-level— Rock permeability determines aquifer potentialsm-03-fault-proximity— Geologic contacts often coincide with fault zonessm-04-liquefaction-risk— Unconsolidated sedimentary geology signals higher liquefaction susceptibility
Classifies land cover at any US location using the National Land Cover Database. Forest, developed, wetlands, cropland—what the satellites see when they look down at this spot.
lat, lng (required); year (optional, defaults to 2021)
NLCD class with code, human-readable name, and broad category
MRLC National Land Cover Database (NLCD). Free, no auth. Status: dev.
Yearly. NLCD releases every 2–3 years—annual cache is fine.
An environmental consultant checking a proposed development site east of Seattle queries t-05 across the parcel boundary. The northern half returns "Deciduous Forest" (NLCD 41), the southern half "Emergent Herbaceous Wetlands" (NLCD 95). That wetland classification triggers a requirement for vt-03-environmental-justice and cv-04-property-regulations to check for protected status before any grading plans are drawn.
f-06-tree-canopy-coverage— Land cover type provides context for canopy percentagecv-01-zoning-lookup— Zoning vs. actual land cover reveals development pressure or conservation gapst-07-wildfire-history— Forested land cover near burn scars suggests fuel load and regrowth statusfa-04-endangered-species— Habitat type drives species occurrence predictions
Retrieves property parcel data including boundaries, ownership, zoning, and assessed value via the Regrid API. The legal shape of the land—which doesn't always match the physical one.
lat, lng (required)
Parcel ID, street address, zoning designation, lot size in square feet
Regrid API. Premium—requires commercial API key. Status: draft.
Monthly. Parcel data updates with county assessor cycles—30-day cache.
A developer scouting lots in Tacoma queries t-06 for a corner parcel and finds it's zoned R-2 (low-density residential) at 7,200 sq ft. Stack with cv-01-zoning-lookup for setback requirements and vl-04-comparable-sales for recent nearby transactions. The parcel's zoning may allow a duplex, but the lot size relative to setbacks is the constraint that matters—a signal only visible when you layer legal shape onto physical shape.
cv-01-zoning-lookup— Zoning designation determines what can be built on the parcelvl-01-property-valuation— Lot size and zoning are primary valuation inputsfm-06-fundus— Parcel geometry enables footprint and setback analysist-02-soil-profile— Soil data within parcel boundaries characterizes buildable area
Retrieves historical wildfire perimeters from the NIFC InterAgency Fire Perimeter History database. Shows fire names, acreage burned, and discovery dates within a configurable search radius. The burn scars that the land remembers even when people forget.
lat, lng (required); radius_km (optional, default 50)
Array of fire events (name, acreage, date), total acreage burned, most recent fire event
NIFC ArcGIS (InterAgency Fire Perimeter History). Free, rate limited.
Hourly during fire season. Active perimeters update frequently—1-hour cache.
A homebuyer looking at property near Sisters, Oregon queries t-07 with a 30km radius and finds 4 fires within the last 20 years, including the 2017 Milli Fire (24,079 acres). The most recent burned within 5km of the property in 2020. Combined with t-05-land-cover (showing dense forest regrowth) and vl-05-insurance-risk, this surfaces a signal that fire insurance costs for this location may be significant—and that defensible space planning should be part of any purchase calculation.
t-05-land-cover— Vegetation type near burn scars indicates regrowth and fuel loadvl-05-insurance-risk— Fire history is a primary driver of insurance premiumsa-05-wind-rose— Prevailing wind direction during fire season determines smoke and fire spread riskt-11-air-quality— Active fires degrade air quality; historical patterns suggest seasonal AQI impacts
Queries the USGS Earthquake Hazards Program to assess seismic activity near any location. Returns earthquake history with magnitude, depth, and distance, plus a computed hazard summary. The Pacific Northwest sits on the Cascadia Subduction Zone—this rill helps quantify what that means for a specific spot.
lat, lng (required); radius_km (default 100); min_magnitude (default 3.0); year_range (default 20)
Array of earthquake events sorted by magnitude, largest magnitude in window, human-readable hazard summary
USGS Earthquake Hazards Program. Free, rate limited.
Daily. Seismic catalogs update continuously—24-hour cache balances freshness with API courtesy.
A structural engineer assessing a retrofit project in downtown Seattle queries t-08 and finds 47 events above M3.0 within 100km over the past 20 years, with the largest at M4.6. The hazard summary reads "moderate seismic activity with frequent minor events." Stack with sm-02-seismic-hazard for probabilistic PGA values and sm-04-liquefaction-risk for soil behavior under shaking. The event history alone is informative; the composition tells you whether the building code assumptions are conservative enough.
sm-02-seismic-hazard— Probabilistic ground motion adds forward-looking risk to historical observationsm-03-fault-proximity— Nearby faults explain the observed seismicity patternsm-04-liquefaction-risk— Soil behavior under shaking amplifies or reduces structural riskvl-05-insurance-risk— Seismic history feeds into hazard insurance calculations
Estimates the Bortle dark-sky class for any location using a heuristic model based on proximity to major metro areas. The difference between seeing 300 stars and 3,000—and a decent proxy for how far you've gotten from the places that never turn off.
lat, lng (required)
Bortle class (1–9), class name (e.g., "Rural sky"), visibility description of what's visible at this level
Local computation (metro-distance heuristic). Free, no API. Status: dev.
Yearly. Light pollution changes with development—but slowly.
An astronomer planning a dark-sky trip from Portland queries t-09 at three candidate sites: Government Camp (Bortle 5, suburban sky), Prineville Reservoir (Bortle 3, rural sky), and Hart Mountain (Bortle 2, pristine dark). Stack with co-10-astrophotography-planner for tonight's best targets and a-10-cloud-forecast for whether you'll actually see them. The Bortle class tells you the ceiling on what's possible; the weather tells you whether tonight's the night.
co-10-astrophotography-planner— Dark-sky quality determines viable imaging targets and exposure timesa-10-cloud-forecast— Clear skies plus dark skies equals a usable observation windowfa-01-bird-watcher— Light pollution disrupts nocturnal bird migration patternssn-05-natural-soundscape— Light and noise pollution often correlate—dark usually means quietcm-01-neighborhood-profile— Light pollution maps inversely to rural character and tranquility
Estimates transportation noise level based on proximity to major airports and highway corridors. Uses a heuristic distance-attenuation model since the DOT noise map provides WMS tiles but no REST API. Not lab-grade, but a useful first pass on how loud a place is.
lat, lng (required)
Estimated dB(A), noise classification (e.g., "Moderate"), nearby noise sources sorted by distance
Local computation (distance attenuation model). Free, no API. Status: dev.
Yearly. Airport and highway locations are stable—annual cache.
A family considering a home near SeaTac airport queries t-10 and finds estimated 68 dB(A) from airport proximity—above the 65 dB threshold where the FAA recommends sound insulation. The noise source list shows the airport at 2.3km and I-5 at 0.8km. Stack with sn-01-ambient-noise for HowLoud's independent assessment and vl-01-property-valuation to check whether noise is already priced into the market. What sounds like a bargain may have a sonic reason.
sn-01-ambient-noise— Cross-validates the heuristic estimate with empirical noise datasn-03-airport-noise— Specific DNL contour data refines the airport componentvl-01-property-valuation— Noise depresses property values—quantifying the discountt-09-light-pollution— Noise and light often co-occur; together they define sensory environment
Reports real-time Air Quality Index from the EPA AirNow network. Shows the dominant pollutant, individual readings for PM2.5, ozone, and other parameters, and health guidance based on the current category. During wildfire season in the PNW, this rill earns its keep.
lat, lng (required)
AQI value (0–500), category name, array of individual pollutant readings with per-pollutant AQI
EPA AirNow API. Free, requires API key.
Hourly. AQI updates with monitoring station reports—1-hour cache.
A runner in Portland checks t-11 on an August morning and finds AQI 165 (Unhealthy), dominant pollutant PM2.5—wildfire smoke from eastern Oregon. The health guidance recommends avoiding prolonged outdoor exertion. Combined with a-05-wind-rose (showing easterly winds) and t-07-wildfire-history (confirming active fires to the east), this surfaces why today's air is bad and how long it might last based on forecast wind shifts.
a-09-pollen-allergens— Pollen plus poor AQI compounds respiratory riskt-07-wildfire-history— Active fires explain sudden AQI spikessa-01-activity-dashboard— AQI thresholds should modify outdoor exercise recommendationsvt-01-air-quality— Vitae's air quality rill provides environmental justice contexta-05-wind-rose— Wind direction predicts smoke movement and clearing timelines
Calculates the magnetic declination for any location using the NOAA NCEI Geomagnetic Calculator, with a WMM 2025 dipole fallback. The difference between where your compass points and where north actually is—a detail that matters more than most people realize.
lat, lng (required); date, altitude (optional)
Declination in degrees (positive = east), direction, annual change rate, inclination, total field intensity in nT
NOAA NCEI Geomagnetic Calculator (primary), WMM 2025 dipole (fallback). Free, no auth.
Yearly. The magnetic field drifts slowly—annual cache.
A surveyor working near Spokane queries t-12 and finds declination of 15.2° East—meaning magnetic north is 15 degrees clockwise from true north. Without this correction, a compass bearing of 0° would actually point N15E, accumulating error over distance. Combined with t-03-elevation-slope for terrain context and c-01-base-terrain for map orientation, this ensures field measurements align with the coordinate system.
c-01-base-terrain— Map orientation requires declination correction for compass navigationt-01-helio-study— Solar azimuth uses true north; compass readings need declination adjustmentki-03-route-planner— Backcountry navigation requires declination-corrected bearingsco-01-night-sky-map— Star chart orientation depends on true-north alignment
atmos (vapor, atmosphere)
What is the sky doing right now, and what has it done before? Atmos covers the atmospheric envelope—weather conditions, forecasts, historical climate patterns, precipitation radar, wind distribution, UV radiation, moon phases, twilight timing, pollen counts, and cloud cover. Most of this data comes from Open-Meteo (free, no auth), with NASA POWER for deep climate history and NOAA for radar imagery. This family provides the temporal dimension that makes static land data come alive.
Real-time weather, multi-day forecasts, historical temperature and precipitation normals, wind direction and speed distributions, UV radiation, pollen concentrations, cloud cover by atmospheric layer, moon phases and illumination, sunrise/sunset and twilight periods, precipitation radar imagery.
Open-Meteo (weather/forecast/UV/wind/pollen/clouds, free, no auth), NASA POWER (deep climate history back to 1981, free), Sunrise-Sunset.org (twilight calculations, free), NOAA WMS (precipitation radar tiles, free), suncalc library (lunar computations, local).
Weather is context for almost everything. A soil profile (t-02) tells you drainage class; combine with a-03-historical-climate and you know how much rain that drainage actually handles. A bloom calendar (f-03) predicts flowering; stack with a-02-seven-day-forecast to know if frost will interrupt it. Wind roses (a-05) explain why wildfire smoke arrives from a specific direction, and why one side of a ridge is wetter than the other. Atmos data turns spatial facts into temporal stories.
Stack a-01-current-conditions + a-06-uv-index + a-09-pollen-allergens + t-11-air-quality + a-10-cloud-forecast for a weekend hike planning check near Mt. Rainier. Current conditions show 62°F with light winds, but UV is "very high" (index 9), birch pollen is moderate, and AQI is 85 (moderate). Cloud forecast shows clearing by 2pm. This surfaces a window: start the hike early, apply sunscreen after the clouds clear, and consider an antihistamine if birch-sensitive. No single data point makes the call; the composition does.
| ID | Name | Description |
|---|---|---|
| a-01 | Current Conditions | Real-time temperature, humidity, pressure, wind, and precipitation |
| a-02 | 7-Day Forecast | Daily highs/lows, precipitation probability, sunrise/sunset, UV index |
| a-03 | Historical Climate | Temperature and precipitation history back to 1981 via NASA POWER |
| a-04 | Precipitation Radar | Live NOAA radar imagery with 6-frame animation (US only) |
| a-05 | Wind Rose | 16-point wind direction and speed distribution from hourly observations |
| a-06 | UV Index | Current and forecast UV radiation with WHO risk classification |
| a-07 | Moon Phase | Lunar phase, illumination, moonrise/moonset via suncalc math |
| a-08 | Astronomical Twilight | Sunrise, sunset, solar noon, and all three twilight periods |
| a-09 | Pollen & Allergens | Pollen counts by species, air quality metrics, 3-day allergy forecast |
| a-10 | Cloud Forecast | Cloud cover by atmospheric layer (low/mid/high) plus visibility |
Real-time weather conditions—temperature, humidity, pressure, wind, and precipitation—for any location worldwide via Open-Meteo. The atmospheric snapshot that anchors everything else to right now.
lat, lng (required)
Temperature (°C), weather description, relative humidity %, wind speed (km/h), barometric pressure (hPa)
Open-Meteo. Free, no auth, rate limited. Status: dev.
Every 30 minutes. Weather updates continuously.
A farmer in the Skagit Valley checks a-01 before deciding whether to irrigate. Temperature is 78°F, humidity 35%, wind 15 km/h from the south. Combined with a-02-seven-day-forecast showing no rain for 5 days and h-05-drought-monitor showing D1 conditions, this surfaces a signal that irrigation is needed now despite recent rains—the evapotranspiration rate is outpacing soil moisture.
a-02-seven-day-forecast— Current conditions provide the baseline that the forecast extendst-11-air-quality— Weather patterns drive pollutant dispersion and AQIcu-01-growing-calendar— Temperature and humidity determine planting and harvest windowssa-01-activity-dashboard— Weather conditions gate outdoor activity recommendations
A full week of daily weather: highs, lows, precipitation probability, sunrise/sunset, wind, and UV. The planning horizon for everything from weekend hikes to construction schedules to bloom predictions.
lat, lng (required)
7-day array: tempMax, tempMin, precipitation probability, sunrise/sunset times, WMO weather code, max wind speed, UV index max
Open-Meteo. Free, no auth, rate limited. Status: dev.
Hourly. Forecasts improve as the target date approaches.
An outdoor event planner in Seattle checks a-02 for a Saturday farmers market setup. The forecast shows 40% precipitation probability with highs of 58°F. Stack with a-04-precipitation-radar the morning of for real-time rain tracking. The 7-day view helps decide whether to rent the tent; the radar on the day helps decide when to set up.
f-03-bloom-calendar— Temperature forecasts predict bloom timing within the weekfg-09-weather-window— Identifies optimal foraging windows within the forecast perioda-04-precipitation-radar— Radar provides the real-time detail that forecasts approximatecu-01-growing-calendar— Frost forecasts trigger protection actions for tender crops
Temperature and precipitation records stretching back to 1981 via NASA POWER, with Open-Meteo for the recent 92 days. The long memory of the atmosphere—climate normals, seasonal patterns, and trend lines that reveal whether this year's weather is typical or exceptional.
lat, lng (required); startDate, endDate (required, YYYY-MM-DD)
Daily observations (tempMean/Max/Min, precipitation), 12-month climate normals (when available), data source indicator
Open-Meteo (recent 92 days) + NASA POWER (deep history). Free, no auth. Status: dev.
Daily. Historical data is stable; recent data updates daily.
A viticulturist evaluating a site in the Willamette Valley pulls a-03 for the past 10 years. Climate normals show an average growing season of 180 days with 2,400 growing degree days—within the range for Pinot Noir. But the trend line reveals warming: the last 5 years average 12% more GDD than the first 5. Combined with t-02-soil-profile (well-drained volcanic soils) and t-03-elevation-slope (south-facing hillside), this surfaces a picture of a site that's becoming more viable for later-ripening varieties than traditional planting guides suggest.
t-02-soil-profile— Historical precipitation explains soil moisture patterns and drainage demandsh-05-drought-monitor— Climate trends contextualize current drought severityf-07-phenology-tracker— Long-term temperature data predicts phenological shift patternscu-08-harvest-predictor— GDD accumulation from historical data calibrates harvest timing models
Live NOAA radar imagery showing rain, snow, and mixed precipitation for US locations. Returns PNG tiles via WMS with a 6-frame animation history at 10-minute intervals—the looping radar you check when the sky looks uncertain.
lat, lng (required, US locations only)
WMS GetMap URL (PNG tile), geographic bounding box, 6 time-indexed frame URLs for animation
NOAA WMS (radar composites). Free, no auth. Status: dev.
Every 5 minutes. Radar scans update continuously.
A construction foreman in Olympia checks a-04 at 6am and sees a precipitation band moving northeast at roughly 20km/h, currently 40km to the southwest. Estimated arrival: 2 hours. Stack with a-02-seven-day-forecast which shows the system clearing by afternoon. The radar says when to tarp the concrete pour; the forecast says when to uncover it.
a-01-current-conditions— Current temp determines whether radar echoes are rain or snowh-07-flood-risk— Intense radar returns over flood-prone areas signal real-time riskc-09-layer-compositor— Radar overlays on terrain maps show precipitation in geographic context
Computes wind direction and speed distribution from Open-Meteo hourly observations, bucketed into a 16-point compass rose with 5 speed tiers. Reveals prevailing winds, calm periods, and seasonal patterns—the invisible architecture of the atmosphere that shapes everything from building design to wildfire behavior.
lat, lng (required); days (optional, default 30)
16-point direction distribution with speed buckets, dominant direction, mean wind speed (km/h)
Open-Meteo (hourly wind data). Free, no auth. Status: dev.
Daily. Wind patterns are statistical—24-hour cache for the distribution.
An architect designing a home on the Oregon coast queries a-05 with 365 days of data. The wind rose shows dominant SW winds at 25-40 km/h from October through March, shifting to NW at 10-20 km/h in summer. This surfaces design implications: the southwest face needs the most weather protection, and natural ventilation strategies should leverage the summer NW breeze. Stack with t-03-elevation-slope to understand whether terrain provides any natural wind shelter.
t-07-wildfire-history— Wind direction during fire season predicts smoke corridorst-11-air-quality— Wind patterns explain AQI variations and pollutant dispersiongr-03-solar-potential— Wind-exposed sites may need different solar mounting strategiessn-01-ambient-noise— Wind direction determines which noise sources are audible
Current UV index and hourly forecast classified into WHO risk levels—low through extreme. Simple, practical, and quietly important: the difference between a pleasant afternoon and a regrettable sunburn.
lat, lng (required)
Current UV index, WHO risk level, today's max UV, 2-day hourly forecast with risk levels
Open-Meteo. Free, no auth. Status: dev.
Every 30 minutes. UV changes with solar angle and cloud cover.
A school planning outdoor field day in Bend, Oregon checks a-06 and finds UV index peaking at 10 ("very high") between 11am and 2pm. The hourly forecast shows it dropping to "moderate" by 3pm. Combined with a-10-cloud-forecast showing afternoon clouds building, this helps schedule the most exposed activities for early morning and late afternoon, with indoor options during the solar peak.
a-10-cloud-forecast— Cloud cover is the primary modulator of ground-level UVt-03-elevation-slope— Elevation increases UV exposure by roughly 10% per 1,000msa-01-activity-dashboard— UV thresholds should modify outdoor exercise scheduling
Current moon phase, illumination percentage, and moonrise/moonset times computed locally via suncalc. No API calls, no rate limits, instant results. Useful for gardening by lunar cycle, astronomy planning, and knowing when the night will be bright or dark.
lat, lng (required); date (optional, defaults to today)
Phase name (e.g., "Waxing Gibbous"), illuminated fraction (0–1), moonrise/moonset times
Local suncalc computation. Free, no API. Status: dev.
Instant. Computed on demand—no caching needed.
A biodynamic gardener near Ashland, Oregon checks a-07 before transplanting. The waning crescent at 12% illumination suggests a root-day in biodynamic practice. Combined with cu-01-growing-calendar (confirming the frost-free window) and t-02-soil-profile (verifying adequate drainage for the intended root crop), the moon phase adds a traditional timing layer to modern soil science.
co-03-moon-phase— Cosmo's moon rill provides extended calendar and quarter datesh-02-tide-tables— Moon phase drives tidal range; spring tides coincide with full/new moonco-10-astrophotography-planner— Moon illumination determines sky brightness for deep-sky imagingcu-01-growing-calendar— Lunar planting calendars use phase data for timing guidance
Sunrise, sunset, solar noon, day length, and all three twilight periods—civil, nautical, and astronomical. The precise boundaries between day and night that photographers call golden hour, astronomers call observing window, and everyone else calls "getting dark."
lat, lng (required); date (optional, defaults to today)
Sunrise/sunset (ISO 8601), day length, civil/nautical/astronomical twilight begin/end times
Sunrise-Sunset.org API. Free, no auth. Status: dev.
Daily. Sun timing changes by a few minutes per day.
A landscape photographer in Olympic National Park checks a-08 for tomorrow. Sunset at 7:42pm, civil twilight ends at 8:14pm—that's the golden hour window. Astronomical twilight ends at 9:48pm, which is when true darkness begins for Milky Way shooting. Combined with a-10-cloud-forecast and t-09-light-pollution (Bortle 2 in the park), this produces a precise shooting schedule for both sunset and astro work.
t-01-helio-study— Solar position adds azimuth (where the sun is) to twilight timing (when)co-10-astrophotography-planner— Astronomical twilight end defines the start of dark-sky observationa-10-cloud-forecast— Clear sky during golden hour is the photographer's compound condition
Current pollen counts for six species (alder, birch, grass, mugwort, olive, ragweed) plus PM10, PM2.5, and dust via the Open-Meteo Air Quality API. Five-tier severity classification and a 3-day hourly forecast for anyone whose body treats spring as a biological event rather than a metaphor.
lat, lng (required)
Pollen array by species with concentration and severity, overall level, European AQI, PM10/PM2.5/dust, 3-day hourly forecast
Open-Meteo Air Quality API. Free, no auth. Status: dev.
Hourly. Pollen counts vary through the day with temperature and wind.
A parent in Portland checks a-09 in late March and finds alder pollen "very high" at 340 grains/m³, grass "low." The 3-day forecast shows alder dropping by Thursday with an incoming rain system. Stack with a-02-seven-day-forecast (confirming rain) and f-03-bloom-calendar (showing alder peak season). This helps plan the week: outdoor activities Wednesday are risky, but Thursday post-rain should be clear. The pollen data alone says "bad today"; the composition says "better Thursday."
t-11-air-quality— AQI plus pollen produces a compound respiratory risk scoref-03-bloom-calendar— Bloom timing predicts which pollen species will be activea-02-seven-day-forecast— Rain washes pollen from the air; forecast predicts relief timingsa-06-mental-wellness— Seasonal allergies correlate with fatigue and mood changes
Cloud cover forecast broken down by atmospheric layer—low, mid, high, and total—plus visibility in meters. Layers can exceed 100% combined because they represent overlapping vertical slices. The difference between "mostly cloudy" and "high thin cirrus over low fog" matters for everything from solar panels to stargazing.
lat, lng (required)
Total cloud cover %, layer breakdown (low/mid/high), visibility in meters, 48-hour hourly forecast
Open-Meteo. Free, no auth, rate limited. Status: dev.
Every 30 minutes. Cloud patterns change rapidly.
An amateur astronomer near Goldendale, Washington checks a-10 and finds current total cloud cover at 80%, but the layer breakdown shows it's all low-level fog (90% low, 5% mid, 0% high). The 48-hour forecast shows low clouds clearing by 10pm as temperatures drop. Stack with co-10-astrophotography-planner and t-09-light-pollution (Bortle 3 at Goldendale Observatory). High clouds would scatter starlight; low fog that clears means tonight may actually be excellent. The layer breakdown changes the conclusion.
co-10-astrophotography-planner— Cloud-free windows determine viable observation timegr-03-solar-potential— Cloud cover directly reduces solar energy productiona-06-uv-index— Clouds modulate UV reaching the grounda-08-astronomical-twilight— Clear sky during twilight creates the best golden hour conditions
hydro (water)
Water defines a landscape as much as the rock beneath it. Hydro covers the full water cycle—stream flow and gage height, tidal predictions, water quality, watershed delineation, drought severity, groundwater levels, flood risk, precipitation accumulation, reservoir storage, and ocean currents. These rills connect to USGS, NOAA, EPA, FEMA, and USBR data streams to answer the question: where is the water, how much is there, and is it safe?
Streamflow and gage height, tidal predictions, water quality parameters (pH, dissolved oxygen, turbidity), watershed boundaries, drought severity indices, groundwater depth, FEMA flood zones and NWS alerts, precipitation accumulation, reservoir storage and capacity, tidal currents and sea surface temperature.
USGS Water Services (streamflow, groundwater, precipitation, free), NOAA CO-OPS (tides, ocean, free), EPA Water Quality Portal (WQI, free), USGS NLDI (watersheds, free), US Drought Monitor (drought, free), FEMA NFHL (flood zones, free), NWS Alerts (warnings, free), USBR RISE (reservoirs, free).
Flood risk alone is a FEMA zone designation. Stack h-07-flood-risk with t-02-soil-profile (drainage class), h-01-stream-gauge (current flow), and a-04-precipitation-radar (incoming rain), and you have a real-time flood intelligence picture that no single dataset provides. Similarly, h-03-water-quality combined with vt-05-toxic-release and h-04-watershed-profile traces water quality issues upstream to their likely sources. Water data gains its real meaning at intersections.
Stack h-04-watershed-profile + h-03-water-quality + h-05-drought-monitor + vt-05-toxic-release + f-06-tree-canopy-coverage for the Tualatin River watershed near Portland. The watershed boundary defines the study area; water quality shows elevated turbidity; drought is D0 (abnormally dry); TRI data shows two reporting facilities upstream; canopy coverage is 38%. This surfaces a signal that reduced flow from drought may be concentrating pollutants—a finding only visible when hydrology, environmental toxicology, and ecology are layered together.
| ID | Name | Description |
|---|---|---|
| h-01 | Stream Gauge Monitor | Real-time streamflow, gage height, and water temperature at nearest USGS station |
| h-02 | Tide Tables | Tidal predictions and observed water levels at nearest NOAA CO-OPS station |
| h-03 | Water Quality Index | Composite WQI from EPA stations (pH, dissolved oxygen, turbidity, conductivity) |
| h-04 | Watershed Profile | Watershed boundaries, drainage area, and stream network via USGS NLDI |
| h-05 | Drought Monitor | Current drought severity (D0–D4) and coverage with weekly trends |
| h-06 | Groundwater Level | Depth to water table at nearest USGS monitoring well |
| h-07 | Flood Risk Map | FEMA flood zones, NWS warnings, and NOAA NWPS forecasts combined |
| h-08 | Precipitation Tracker | Rainfall totals (24h and 7d) at nearest USGS rain gauge |
| h-09 | Lake & Reservoir Status | Water storage, capacity, and elevation for western US reservoirs |
| h-10 | Ocean Current Tracker | Tidal currents, sea surface temperature, and salinity at NOAA stations |
Real-time streamflow, gage height, and water temperature at the nearest USGS monitoring station, with flood stage categories from NOAA. The pulse of every river and creek that has a gauge on it.
lat, lng (required)
Discharge (cfs), gage height (ft), water temp, USGS site info, flood stage categories (action/minor/moderate/major)
USGS Water Services + NOAA NWPS. Free, rate limited.
Every 15 minutes. USGS gauges report at 15-minute intervals.
A kayaker checks h-01 for the Clackamas River before a weekend trip. Current discharge is 2,400 cfs at gage height 4.2 ft—below the 7.0 ft action stage. Stack with a-04-precipitation-radar showing rain moving in and a-02-seven-day-forecast calling for 2 inches by Saturday. The river is runnable now, but may approach action stage by Sunday afternoon—a signal to plan the put-in for Saturday morning.
h-07-flood-risk— Real-time gage height feeds directly into flood risk assessmenth-04-watershed-profile— Watershed area determines how quickly precipitation translates to streamflowa-04-precipitation-radar— Incoming precipitation predicts gage height changes
Tidal predictions (high/low times and heights) and observed water levels at the nearest NOAA CO-OPS station. The rhythm of the coast that every beachcomber, clam digger, and boat operator needs to know.
lat, lng (required)
High/low tide predictions (times, heights), current observed level, station metadata
NOAA CO-OPS. Free, rate limited.
Daily. Predictions are precomputed; observed levels update in real-time.
A razor clam digger on the Long Beach Peninsula checks h-02 and finds a -1.2 ft low tide at 6:47am—excellent for digging. Stack with a-07-moon-phase (new moon, which means spring tides and the lowest lows) and a-01-current-conditions (checking for wind and rain). The minus tide plus new moon is the compound condition that makes the best clamming days.
a-07-moon-phase— Lunar phase determines spring vs. neap tide rangeh-10-ocean-current— Current direction affects water quality at the tide stationfg-06-seaweed-coastal— Low tides expose intertidal foraging zones
Water quality assessment from the EPA Water Quality Portal with a composite WQI computed from pH, dissolved oxygen, turbidity, and conductivity. Not every stream is drinkable, and this rill surfaces the ones that warrant a closer look.
lat, lng (required)
Composite WQI score, individual parameter readings, monitoring station details
EPA Water Quality Portal. Free, no rate limit.
Weekly. Monitoring data updates with sampling schedules.
A community group monitoring Johnson Creek in SE Portland queries h-03 and finds WQI 62 (marginal), with elevated turbidity after recent rains. Stack with h-04-watershed-profile to trace the drainage area and vt-05-toxic-release to check for upstream pollution sources. The WQI number is a symptom; the composition helps diagnose causes.
h-04-watershed-profile— Watershed boundaries show what drains into the monitoring pointvt-05-toxic-release— TRI facilities upstream may explain water quality degradationvt-02-water-quality— Vitae's water quality rill adds environmental justice context
Delineates the watershed boundary, drainage area, and stream network for any point in the contiguous US using the USGS NLDI. Everything upstream flows to this point—understanding the catchment is understanding the water.
lat, lng (required)
Basin polygon (GeoJSON), COMID identifier, flowline network (GeoJSON), connected USGS sites
USGS NLDI (Network Linked Data Index). Free, no rate limit.
Yearly. Watershed boundaries are essentially static.
An environmental planner studying a proposed gravel mine near Issaquah queries h-04 to delineate the watershed. The basin boundary shows that runoff from the site flows directly into Issaquah Creek, which connects to Lake Sammamish. Stack with h-03-water-quality for current creek conditions and fa-04-endangered-species for salmon run data. The watershed profile turns a point-location proposal into a hydrological impact area—a spatial context no other rill provides.
h-03-water-quality— Water quality within the watershed provides baseline conditionsh-01-stream-gauge— Connected USGS sites monitor flow throughout the watershedc-09-layer-compositor— Watershed overlays on terrain maps show drainage patterns visuallyt-05-land-cover— Land use within the watershed drives runoff characteristics
Current drought severity on the D0–D4 scale from the US Drought Monitor, with DSCI (Drought Severity and Coverage Index) and weekly trend data. The slow-moving water crisis that doesn't make the news until it's too late to do much about it.
lat, lng (required)
Drought severity level, DSCI score, weekly trend (improving/worsening/stable)
US Drought Monitor + FCC Census API (for county lookup). Free, no rate limit.
Weekly. The Drought Monitor updates every Thursday.
A rancher in central Oregon checks h-05 and finds D2 (Severe Drought) with a worsening trend over 4 weeks. Stack with h-09-lake-reservoir (showing Prineville Reservoir at 45% capacity) and a-03-historical-climate (showing below-normal precipitation for 6 consecutive months). This surfaces a signal that irrigation water allocations may be cut—a planning horizon measured in weeks, not days.
h-09-lake-reservoir— Reservoir levels provide the storage context for drought severitya-03-historical-climate— Precipitation deficit over time explains drought trajectorycu-08-harvest-predictor— Drought modifies crop yield predictionst-07-wildfire-history— Drought increases wildfire risk; historical fires cluster in drought years
Groundwater levels and depth to water table at the nearest USGS monitoring well. The invisible water supply that millions of rural homes and farms depend on—measured in feet below the surface.
lat, lng (required)
Depth to water table (gauge reading), monitoring well metadata
USGS Water Services (parameter code 72019). Free, rate limited.
Every 15 minutes. Groundwater levels change slowly but are monitored continuously.
A prospective homebuyer evaluating a well-dependent property in rural Thurston County queries h-06 and finds the water table at 42 feet below surface at the nearest monitoring well. Stack with t-04-geology-bedrock (glacial outwash deposits—good aquifer material) and h-05-drought-monitor (no current drought). The 42-foot depth is normal for the area, and the geology supports good well yield. But the drought monitor's trend line is worth watching.
t-04-geology-bedrock— Rock type determines aquifer capacity and recharge rateh-05-drought-monitor— Drought depletes aquifers over time; depth trends reveal the impactt-02-soil-profile— Soil permeability determines how quickly rain recharges groundwater
Combines FEMA flood zone designations, NWS flood warnings, and NOAA NWPS forecasts into a unified flood risk assessment. The compound rill that answers "will this flood?" by triangulating regulatory maps, real-time alerts, and river forecasts.
lat, lng (required)
FEMA flood zone designation, computed risk level, array of active flood alerts
FEMA NFHL + NWS Alerts + NOAA NWPS. Free, rate limited. Compound source.
Every 5 minutes during active events. FEMA zones are static; alerts update in real-time.
A homebuyer evaluating a property along the Snoqualmie River queries h-07 and finds FEMA Zone AE (1% annual chance flood plain) with a Base Flood Elevation of 32 ft. No current NWS alerts. Stack with vl-02-flood-risk for insurance cost implications and h-01-stream-gauge for current river level relative to flood stage. The FEMA zone means mandatory flood insurance for any federally-backed mortgage—a financial signal that the map makes visible.
vl-02-flood-risk— Insurance cost quantifies the financial impact of the flood zoneh-01-stream-gauge— Current gage height shows proximity to flood stage right nowt-02-soil-profile— Poor drainage compounds flood risk in low-lying areasvl-01-property-valuation— Flood zone designation depresses property values
Rainfall totals at the nearest USGS rain gauge—24-hour and 7-day accumulation in inches. The simple question of "how much has it rained?" that turns out to drive half the other questions in this domain.
lat, lng (required)
24-hour total (inches), 7-day total (inches), station metadata
USGS Water Services (parameter code 00045). Free, rate limited.
Every 15 minutes. Rain gauges report at standard USGS intervals.
A construction manager checks h-08 after a November atmospheric river hits the Puget Sound region. The 7-day total reads 4.8 inches—well above the monthly average of 5.4 inches, delivered in one week. Stack with h-07-flood-risk (flood warnings now active) and t-02-soil-profile (clay soils saturated). This confirms what the mud on the job site already suggests: no grading work this week.
h-07-flood-risk— Precipitation total is the primary driver of flood eventsh-01-stream-gauge— Rain gauge data correlates with stream gage response after a lagt-02-soil-profile— Saturated soils after heavy rain signal erosion and stability risks
Water storage, capacity, and elevation at the nearest reservoir—primarily via USBR RISE for western US dams, with USGS fallback for other regions. The water supply gauge that measures whether the summer will be wet or dry for irrigators, municipalities, and fish.
lat, lng (required)
Current storage (acre-feet), surface elevation, reservoir metadata (name, capacity, dam info)
USBR RISE (primary, western US) + USGS Water Services (fallback). Free, no rate limit.
Daily. Reservoir levels update once or twice per day.
A municipal water planner in Bend checks h-09 for Wickiup Reservoir in early July. Storage is at 38% of capacity with elevation dropping 0.5 ft/week. Stack with h-05-drought-monitor (D1 drought) and a-03-historical-climate (below-normal snowpack last winter). This surfaces a signal that August water restrictions may be necessary—the reservoir drawdown rate plus the drought trajectory suggests the math doesn't work for a normal irrigation season.
h-05-drought-monitor— Drought severity contextualizes storage levelsa-03-historical-climate— Snowpack and precipitation determine refill potentialfa-04-endangered-species— Minimum pool levels affect fish habitat downstream
Tidal currents, sea surface temperature, and salinity at the nearest NOAA CO-OPS station. V1 covers tidal currents from CO-OPS; global ocean current data is planned for v2. The marine complement to the freshwater hydrology rills.
lat, lng (required)
Current speed/direction, water temperature, salinity, station metadata
NOAA CO-OPS. Free, rate limited. Status: dev.
Every 6 minutes. CO-OPS stations report at high frequency.
A sea kayaker planning a crossing of Admiralty Inlet in Puget Sound queries h-10 and finds current speed of 2.1 knots flowing south on the ebb. Stack with h-02-tide-tables (slack tide at 3:15pm) and a-05-wind-rose (afternoon northerly wind opposing the ebb current—which creates standing waves). The safe crossing window is around slack tide; the current and wind combination at other times is the hazard that this composition reveals.
h-02-tide-tables— Tidal predictions determine current strength and direction timinga-05-wind-rose— Wind opposing current creates hazardous water conditionsfa-06-marine-life— Temperature and salinity drive marine species distribution
seismos (earthquake, tremor)
The Pacific Northwest sits on the Cascadia Subduction Zone, making seismic awareness more than academic. Seismos is the dedicated family for earthquake intelligence—recent events, probabilistic hazard scores, fault proximity, liquefaction susceptibility, and long-term seismicity patterns. While t-08-seismic-risk in terra provides a quick seismic snapshot, seismos goes deeper: NSHMP probabilistic ground motion, mapped Quaternary faults, California-specific liquefaction zones, and magnitude-frequency analysis spanning decades.
Real-time earthquake events, probabilistic seismic hazard (PGA, spectral acceleration), Quaternary fault mapping (slip rate, type, last activity), soil liquefaction susceptibility, and historical earthquake catalogs with magnitude distribution analysis.
USGS GeoJSON Feeds (real-time earthquakes, free), USGS NSHMP (probabilistic hazard, free), USGS Quaternary Faults (fault database, free), CGS Earthquake Hazard Zones (California liquefaction, free), USGS FDSN Event API (historical catalog, free). All status: dev.
A seismic hazard score alone is a number. Stack it with t-04-geology-bedrock (soft sediments amplify shaking), t-02-soil-profile (clay soils liquefy), and vl-05-insurance-risk (earthquake insurance pricing), and you have a structural risk profile that no single data source provides. Seismos rills connect geology to risk to financial impact—the full chain from cause to consequence.
Stack sm-02-seismic-hazard + sm-03-fault-proximity + sm-04-liquefaction-risk + t-04-geology-bedrock + t-02-soil-profile for a proposed building site in Seattle's SoDo district. The hazard score is 72/100; the nearest fault is the Seattle Fault at 1.2km; liquefaction susceptibility is "high" on the alluvial fill; bedrock is Quaternary glacial deposits. This surfaces a signal that the site requires deep foundation engineering—a finding that only the full seismic-geologic composition makes clear.
| ID | Name | Description |
|---|---|---|
| sm-01 | Recent Earthquakes | Nearby earthquakes from the past 30 days via USGS real-time feeds |
| sm-02 | Seismic Hazard Score | 0–100 hazard index from USGS NSHMP probabilistic ground motion |
| sm-03 | Fault Line Proximity | Nearest Quaternary faults with slip rate, type, and age of last activity |
| sm-04 | Liquefaction Risk | Soil liquefaction susceptibility via CGS zones or USGS Vs30 proxy |
| sm-05 | Historical Seismicity | Earthquake catalog with magnitude distribution and temporal patterns |
Nearby earthquakes from the past 30 days via USGS real-time GeoJSON feeds. Filters by configurable radius, shows magnitude, depth, distance, and time. When the ground shakes, this is the first rill to check.
lat, lng (required); radiusKm (optional, default 250)
Array of earthquake events, total count, maximum magnitude in period
USGS GeoJSON Feeds. Free, rate limited. Status: dev.
Every 60 seconds. Real-time feeds update with new detections.
After feeling a jolt in Portland, a resident queries sm-01 within 100km and finds a M3.2 event 12 minutes ago, 45km SSE at 8km depth. Combined with sm-03-fault-proximity, the event falls near the mapped East Bank Fault. Not dangerous, but a reminder that the region is seismically active—and a prompt to check sm-02-seismic-hazard for the probabilistic picture.
sm-03-fault-proximity— Locating events relative to mapped faults reveals seismic source patternssm-05-historical-seismicity— Recent events in the context of decades of historyk-01-anomaly-sentinel— Seismic events can trigger automated alert cascades
A 0–100 seismic hazard index computed from USGS NSHMP probabilistic ground motion data. Returns Peak Ground Acceleration and spectral accelerations normalized into a human-readable risk score. The forward-looking counterpart to historical earthquake counts.
lat, lng (required); vs30 (optional, default 760 m/s for firm rock)
Hazard score (0–100), PGA value, hazard level classification
USGS NSHMP (National Seismic Hazard Model). Free, no rate limit. Status: dev.
Yearly. NSHMP updates with major model revisions.
A structural engineer in Olympia queries sm-02 for a hospital retrofit project and finds hazard score 78/100 with PGA 0.55g at 2% probability in 50 years. The vs30 at the site is 280 m/s (soft soil), which amplifies the already high base hazard. Stack with sm-04-liquefaction-risk and the foundation design requirements become clear: this site needs seismic isolation or deep piles, not shallow spread footings.
sm-03-fault-proximity— Nearby faults explain why the hazard score is elevatedsm-04-liquefaction-risk— Soft-soil sites with high hazard scores have compound riskvl-05-insurance-risk— Hazard score correlates with earthquake insurance pricing
Identifies the nearest mapped Quaternary faults from the USGS Fault and Fold Database—fault name, distance, slip rate, type, and age of last activity. The cracks in the earth that are still moving, even if slowly.
lat, lng (required); radiusM (optional, default 50,000m)
Array of fault features (name, distance, slip rate, type), nearest fault, total count
USGS Quaternary Faults ArcGIS service. Free, rate limited. Status: dev.
Yearly. Fault databases update with new mapping campaigns.
A developer scouting a site in the East Bay hills near Oakland queries sm-03 and finds the Hayward Fault at 0.8km with a slip rate of 9 mm/yr—one of the most hazardous faults in the US. The Calaveras Fault is at 12km. Combined with sm-02-seismic-hazard (hazard score 85) and sm-04-liquefaction-risk, the composition makes clear that this site requires the most stringent seismic engineering. The fault proximity alone is the signal; the composition defines the response.
sm-02-seismic-hazard— Fault proximity drives the probabilistic hazard calculationt-04-geology-bedrock— Geologic contacts often align with fault tracessm-01-recent-earthquakes— Events clustered near faults confirm active seismicity
Assesses soil liquefaction susceptibility—where the ground might behave like a liquid during shaking. Uses CGS regulatory zones for California; estimates via USGS Vs30 shear-wave velocity elsewhere. The hidden hazard beneath seemingly solid ground.
lat, lng (required)
In-zone status (boolean), susceptibility level, data source indicator
CGS Earthquake Hazard Zones (CA) + USGS NSHMP Vs30 (elsewhere). Free. Status: dev.
Yearly. Liquefaction maps update with new geotechnical data.
A homebuyer considering a property in San Francisco's Marina District queries sm-04 and finds it inside a CGS-designated liquefaction zone—the same area that saw dramatic damage in 1989's Loma Prieta earthquake. Stack with sm-02-seismic-hazard (high) and t-04-geology-bedrock (artificial fill over bay mud). The composition tells a clear story: this is beautiful real estate on fundamentally unstable ground. The disclosure is required; the understanding is optional but valuable.
sm-02-seismic-hazard— High hazard on liquefiable soil is the worst-case combinationt-02-soil-profile— Sandy, saturated soils are most susceptible to liquefactiont-04-geology-bedrock— Fill and alluvial deposits signal liquefaction potentialvl-05-insurance-risk— Liquefaction zone status affects insurance availability and cost
The earthquake record spanning decades—magnitude distribution, temporal patterns, and event density from the USGS FDSN Event API. Where sm-01 shows what happened this month, this rill shows what's been happening for the last 20 years.
lat, lng (required); radiusKm (default 100); years (default 20)
Event array, total count, magnitude distribution buckets, largest event
USGS FDSN Event API. Free, rate limited. Status: dev.
Daily. Historical catalog updates as events are reviewed and relocated.
A seismologist analyzing the Cascadia region pulls sm-05 for 50 years within 200km of Seattle. The magnitude distribution follows the expected Gutenberg-Richter curve, with 2,400 events above M2.0 and 12 above M5.0. The temporal pattern shows clusters in the 2001 Nisqually sequence. Combined with sm-02-seismic-hazard, this historical record validates the probabilistic model—or reveals where it might be conservative or optimistic.
sm-01-recent-earthquakes— Recent events in the context of long-term patternssm-02-seismic-hazard— Historical rates validate or challenge probabilistic modelssm-03-fault-proximity— Event clusters mapped to known faults reveal source characteristics
sonic (sound)
How loud is this place, and what makes it that way? Sonic maps the auditory environment—ambient noise levels, noise source identification, airport DNL contours, traffic noise estimates, and natural soundscape health. This is the sensory data that real estate listings never mention but that profoundly affects livability. A Bortle 2 dark sky with a 75 dB traffic drone is not the retreat it appears to be on paper.
Composite ambient noise levels (dBA), noise source mapping (traffic, aviation, rail, local), FAA airport DNL contours, road noise distance-attenuation estimates, natural soundscape health indices (NDSI proxy from NPS data).
HowLoud Soundscore API (ambient noise, key required), BTS National Transportation Noise Map (traffic/aviation noise, free), NPS Geospatial Sound Model (natural soundscape, free). SN-01 and SN-02 require auth; SN-03 through SN-05 are free.
Noise data transforms property analysis. Stack sn-01-ambient-noise with vl-01-property-valuation and you can quantify the noise discount on property values. Combine sn-05-natural-soundscape with t-09-light-pollution and fa-01-bird-watcher for a sensory environment assessment that captures what makes a place feel like wilderness vs. suburb. Sound is the invisible dimension of place that most data platforms ignore.
| ID | Name | Description |
|---|---|---|
| sn-01 | Ambient Noise Level | Composite noise estimate (dBA) from HowLoud and BTS data |
| sn-02 | Noise Source Map | Identifies and classifies nearby noise sources (traffic, aviation, rail) |
| sn-03 | Airport Noise Contour | DNL aviation noise contours at 65, 70, and 75 dB thresholds |
| sn-04 | Traffic Noise Estimate | Road noise from BTS data with distance attenuation model |
| sn-05 | Natural Soundscape Index | Soundscape health via NPS model comparing natural to human-impacted levels |
Estimated ambient noise level in dB(A) combining HowLoud Soundscore and BTS National Transportation Noise Map data. The single-number answer to "how loud is this place?"—with source breakdown to explain why.
lat, lng (required)
Estimated dB(A), noise category, HowLoud Soundscore (if available), BTS noise levels by transport mode
HowLoud API (requires key) + BTS Noise Map (free). Status: dev.
Every 90 days. Ambient noise changes with infrastructure, not weather.
A homebuyer comparing two neighborhoods in Portland queries sn-01 for both. Sellwood returns 48 dB(A) ("quiet residential"); Foster-Powell returns 62 dB(A) ("moderate") due to I-205 proximity. The 14 dB difference represents roughly a 4x perceived loudness increase. Stack with vl-01-property-valuation—the Foster-Powell property is $80K cheaper, and now you know part of the reason.
vl-01-property-valuation— Noise levels inversely correlate with property valuest-09-light-pollution— Light and noise together define the sensory environmentcm-01-neighborhood-profile— Noise is a key component of neighborhood character
Identifies and maps nearby noise sources—traffic, flight paths, rail lines, and local sources—with dB contributions from each. The diagnostic companion to sn-01: not just how loud, but what's making it loud.
lat, lng (required)
Array of noise sources with type, dB contribution, and distance; dominant source; total composite dB
HowLoud API + BTS Noise Map. Auth required. Status: dev.
Every 90 days. Source inventory changes slowly.
A noise-sensitive buyer finds a seemingly quiet property near Olympia, but sn-02 reveals the dominant source is JBLM (Joint Base Lewis-McChord) military flight operations at 58 dB—intermittent but loud when active. Traffic noise is only 42 dB. The ambient average hides the peak events. Stack with sn-03-airport-noise for formal contour data and a-05-wind-rose to check whether prevailing winds carry sound toward or away from the site.
sn-01-ambient-noise— Source map explains the composite noise levelsn-03-airport-noise— Formal contour data for aviation noise sourcesa-05-wind-rose— Wind direction modulates how far noise carries
DNL (Day-Night Level) aviation noise contours from the BTS National Transportation Noise Map. Shows whether a location falls within the 65, 70, or 75 dB DNL zones where the FAA recommends or requires sound insulation.
lat, lng (required)
DNL level, contour band, inside-contour status, impact summary
BTS National Transportation Noise Map. Free, rate limited. Status: dev.
Yearly. Contour maps update with airport operations changes.
A developer evaluating a mixed-use project near SeaTac queries sn-03 and finds DNL 68 dB—above the 65 dB threshold. This triggers FAA Part 150 compatibility guidelines recommending sound insulation for residential use. Combined with cv-01-zoning-lookup (zoned for mixed-use) and vl-01-property-valuation, the noise contour doesn't prohibit the project but adds insulation costs that affect the pro forma.
cv-01-zoning-lookup— Some jurisdictions restrict residential development in high-DNL zonesvl-01-property-valuation— Airport noise contours depress property values within the zonesn-01-ambient-noise— DNL is a 24-hour average; ambient noise shows the real-time picture
Road traffic noise estimate using BTS road noise data and a distance attenuation model. The highway you can't see might still be the loudest thing in your life if the wind is right and the terrain is flat.
lat, lng (required)
Estimated dB(A), nearest roads with type and distance, estimation method
BTS National Transportation Noise Map. Free, rate limited. Status: dev.
Yearly. Road networks and traffic volumes change slowly.
A property buyer checks sn-04 for a rural-feeling lot on Whidbey Island and finds estimated 52 dB from SR-20 at 600m. The road is screened by trees and not visible from the lot, but sound carries. Stack with t-03-elevation-slope—the lot sits below road grade, which means terrain partially blocks sound. The 52 dB estimate may be conservative in this case. Terrain and vegetation modify noise in ways the flat-terrain model doesn't capture.
sn-01-ambient-noise— Cross-validates the traffic-specific estimate with composite noise datat-03-elevation-slope— Terrain between source and receiver modifies sound propagationt-10-noise-profile— Terra's noise heuristic provides an independent estimate for comparison
Natural soundscape health index from the NPS Geospatial Sound Model. Compares natural ambient sound to existing (human-impacted) levels, computing a proxy NDSI that measures how much of the soundscape is birds and wind vs. engines and sirens.
lat, lng (required)
NDSI proxy, NPS sound levels (natural vs. existing), soundscape ratio, category classification
NPS Geospatial Sound Model. Free, computed locally. Status: dev.
Yearly. NPS sound models update with new field measurements.
An ecologist studying habitat quality in the Gifford Pinchot National Forest queries sn-05 and finds NDSI 0.82 (excellent natural soundscape) with existing levels only 3 dB above natural baseline. Stack with t-09-light-pollution (Bortle 2), fa-01-bird-watcher (high species diversity), and f-05-ecoregion-profile. This composition paints a picture of a site where both visual and acoustic wilderness is intact—valuable for conservation prioritization and as a control site for impact studies.
t-09-light-pollution— Dark and quiet together define wilderness qualityfa-01-bird-watcher— Bird diversity correlates with soundscape healthf-05-ecoregion-profile— Ecological context explains what the natural soundscape should sound likecm-01-neighborhood-profile— Soundscape health measures what demographics can't: sensory quality of place
cosmo (universe, order)
Look up. Cosmo covers the sky beyond the atmosphere—star charts, planet positions, moon phases, meteor showers, ISS passes, aurora forecasts, solar activity, deep sky objects, satellite tracking, and astrophotography planning. Most of these rills run on the astronomy-engine library with zero API calls, making them fast and free. This is the family for anyone who looks at the night sky and wants to know what they're seeing, or plans to see it.
Star and constellation positions, planetary ephemerides, lunar phase and calendar, meteor shower predictions, ISS orbital passes, aurora probability and Kp index, solar flares and CME tracking, Messier catalog deep sky objects, satellite pass predictions, and astrophotography target planning with dark-sky windows.
astronomy-engine (local computation, free), N2YO API (ISS/satellite tracking, key required), NOAA/SWPC (aurora/solar activity, free), NASA DONKI (solar events, key required), NASA APOD (picture of the day, key required), static IMO meteor shower catalog (free). CO-05, CO-07, CO-09, CO-10 require API keys; the rest are free.
Cosmo data transforms a location from a point on a map into a point under a sky. Stack co-10-astrophotography-planner with t-09-light-pollution, a-10-cloud-forecast, and a-07-moon-phase for a complete "will I see anything tonight?" assessment. Combine co-06-aurora-forecast with a-01-current-conditions (clear skies?) and ki-03-route-planner (drive time to a dark site) for an aurora chase plan. The celestial data is global; composition grounds it in a specific time and place.
Stack co-10-astrophotography-planner + t-09-light-pollution + a-10-cloud-forecast + a-08-astronomical-twilight + co-08-deep-sky-objects for a planned trip to Cherry Springs, Pennsylvania (Bortle 2). Astronomical twilight ends at 9:48pm. Cloud forecast shows clearing by 10pm. Moon sets at 11:30pm (waxing crescent, 22% illumination). Best deep-sky targets: M31 (Andromeda Galaxy) transiting at midnight, M42 (Orion Nebula) rising at 10pm. The composition produces a minute-by-minute shooting schedule that no single rill could construct.
| ID | Name | Description |
|---|---|---|
| co-01 | Night Sky Map | Interactive star chart with constellations, planets, and deep sky objects |
| co-02 | Planet Tracker | All visible planets with positions, rise/set times, and magnitude |
| co-03 | Moon Phase & Calendar | Lunar phase, illumination, moonrise/set, and upcoming quarter dates |
| co-04 | Meteor Shower Calendar | Upcoming showers with peak dates, ZHR, and viewing condition scores |
| co-05 | ISS Tracker | Real-time ISS position and visible pass predictions via N2YO |
| co-06 | Aurora Forecast | Kp index, OVATION aurora probability, and geomagnetic storm level |
| co-07 | Solar Activity Monitor | Sunspot count, solar flares, CME tracking, and solar wind speed |
| co-08 | Deep Sky Object Guide | Galaxies, nebulae, and clusters visible tonight with viewing details |
| co-09 | Satellite Tracker | Visible satellite passes including Starlink with timing and brightness |
| co-10 | Astrophotography Planner | Best targets tonight with dark windows, moon interference, and APOD |
An interactive star chart showing constellation outlines, planets, bright stars, and deep sky objects above the horizon right now. All computed client-side via astronomy-engine—your personal planetarium with no server calls.
lat, lng (required); date (optional)
Visible celestial positions, planet positions, bright star list with magnitude and constellation
astronomy-engine (local computation). Free, no API. Status: dev.
Every 5 minutes. The sky rotates; positions need frequent updates for interactivity.
A family camping at Cape Lookout on the Oregon coast opens co-01 at 10pm. The chart shows Jupiter blazing at magnitude -2.5 in the southwest, the Summer Triangle overhead, and the Milky Way arcing from Sagittarius to Cassiopeia. A child asks "what's that bright one?" and the star chart identifies it instantly. Stack with t-09-light-pollution (Bortle 4 at this coastal site) to understand which objects are actually visible vs. theoretically above the horizon.
t-09-light-pollution— Bortle class determines the limiting magnitude for naked-eye objectsa-10-cloud-forecast— Cloud cover determines what's visible regardless of what's above the horizonco-08-deep-sky-objects— Adds telescope targets to the naked-eye star chart
Tracks all visible planets with real-time positions, rise/set times, visual magnitude, and elongation. Zero API calls—all astronomy-engine computation. Answers the question: which planets can I see tonight, and when?
lat, lng (required); date (optional)
Array of planet data (position, rise/set, magnitude, elongation), currently visible subset
astronomy-engine. Free, local. Status: dev.
Every 5 minutes. Planet positions shift noticeably over an evening.
An amateur astronomer in Bellingham checks co-02 before setting up a telescope. Saturn is at magnitude 0.8, rising at 8:15pm in the east at elongation 160° (nearly opposition—best viewing of the year). Mars rises at 11:30pm. The rise times help schedule the evening: start with Saturn, switch to Mars after midnight. Stack with a-10-cloud-forecast and a-08-astronomical-twilight for the full timing picture.
co-01-night-sky-map— Planet positions plotted on the star chart for visual referenceco-10-astrophotography-planner— Planets near opposition are prime imaging targetsa-08-astronomical-twilight— Twilight timing determines when planets become visible
Current lunar phase, illumination, moonrise/set, and upcoming quarter dates via astronomy-engine. The atmos family's a-07 covers the basics; this rill adds the extended calendar and upcoming quarter predictions that planners need.
lat, lng (required); date (optional)
Phase angle, phase name, illumination, rise/set times, next quarter dates (new/first/full/third)
astronomy-engine. Free, local. Status: dev.
Every 10 minutes. Moon position changes perceptibly over an hour.
A tide pool photographer on the Olympic coast checks co-03 for the next two weeks. New moon is in 3 days (spring tides for extreme low exposures), full moon in 17 days (bright nights but weaker lows). Stack with h-02-tide-tables for the exact low tide windows and a-10-cloud-forecast for the weather outlook. The moon calendar drives the tide schedule which drives the access to intertidal zones.
h-02-tide-tables— Lunar quarters directly determine spring and neap tide magnitudesco-04-meteor-shower— Moon illumination during a meteor shower determines visibilityco-10-astrophotography-planner— Moonless nights are required for deep-sky imaging
Upcoming meteor showers with peak dates, zenithal hourly rates, radiant positions, and viewing condition scores. Uses a static IMO-derived dataset with astronomy-engine for radiant altitude computation. The annual rhythm of the sky's free fireworks.
lat, lng (required); date (optional)
Shower array (name, peak date, ZHR, radiant position, viewing score), active showers now, next peak
Static IMO catalog + astronomy-engine. Free. Status: dev.
Daily. Shower list is static; radiant altitude changes nightly.
An educator planning an August star party near Mt. Hood checks co-04 and finds the Perseids peaking on August 12 with ZHR 100. The radiant rises above 30° by 11pm at this latitude. Viewing score: 85/100 (high radiant, moderate moon interference). Stack with co-03-moon-phase (waxing crescent, 35% illumination—sets by midnight) and t-09-light-pollution (Bortle 3 at the site). After moonset, conditions should be excellent for 3–4 hours of dark sky.
co-03-moon-phase— Moon illumination and set time determine usable dark hourst-09-light-pollution— Bortle class sets the floor for meteor visibilitya-10-cloud-forecast— Clear sky is the non-negotiable condition for meteor watching
Real-time International Space Station position and upcoming visible pass predictions via N2YO. When will it fly over your head, and will it be bright enough to see? Also lists the current crew, because knowing who's up there makes watching it more interesting.
lat, lng (required)
Current ISS position, visible pass predictions (time, duration, max elevation, magnitude), crew manifest
N2YO API. Requires API key. Status: dev.
Every 6 hours. Pass predictions are stable for days; position updates in real-time.
A parent in Seattle checks co-05 and finds a bright ISS pass tomorrow at 9:12pm—magnitude -3.8, max elevation 72°, duration 6 minutes, entering from the WSW. That's brighter than any star and nearly overhead. Stack with a-10-cloud-forecast (partly cloudy but the western sky is clear) and a-08-astronomical-twilight (nautical twilight, sky dark enough to see it). A perfect evening activity that needs exactly this kind of multi-rill timing.
a-10-cloud-forecast— Clear sky in the pass direction is required for visibilitya-08-astronomical-twilight— ISS is visible only during twilight when it's sunlit but the sky is darkco-09-satellite-tracker— ISS is one of many satellites; the tracker shows the broader picture
Kp geomagnetic index, OVATION aurora probability for the observer's location, and NOAA geomagnetic storm level. The northern lights are occasionally visible from the northern PNW during strong storms—this rill tells you when to look.
lat, lng (required)
Kp index, aurora probability (%) at observer location, storm level, viewline latitude
NOAA/SWPC public JSON feeds. Free, no auth. Status: dev.
Every 5 minutes. Geomagnetic conditions change rapidly during storms.
A photographer in Bellingham (48.75°N) checks co-06 after hearing about a CME impact. Kp is 7 (G3 storm), OVATION probability 45% at this latitude. Stack with a-10-cloud-forecast (clearing after 10pm), t-09-light-pollution (Bortle 5 in town, Bortle 3 at Baker Lake), and ki-03-route-planner (45-minute drive to Baker Lake). The composition produces a chase plan: drive north, arrive by 10:30pm, face north over the lake for the darkest horizon. Kp 7 at 48°N is uncommon but real.
co-07-solar-activity— Solar flares and CMEs drive the geomagnetic storms that cause auroraa-10-cloud-forecast— Clear northern horizon is essential for aurora viewingt-09-light-pollution— Dark sites make faint aurora visible that bright locations wash out
Monitors sunspot count, solar flare alerts, CME tracking, solar wind speed, and flare probability forecasts. Combines NOAA/SWPC feeds with NASA DONKI event data. The upstream driver for aurora, and a useful signal for anyone running satellite-dependent infrastructure.
None (global data, not location-specific)
Sunspot number, recent flares array, flare probability forecasts, solar wind speed/density
NOAA/SWPC + NASA DONKI. DONKI requires NASA API key. Status: dev.
Every 30 minutes. Solar activity changes over hours to days.
A network administrator managing satellite uplinks queries co-07 and sees an X2.1 flare with an associated Earth-directed CME, ETA 36 hours. Solar wind speed is elevated at 650 km/s. Stack with co-06-aurora-forecast for the geomagnetic impact prediction and re-08-connection-quality for current satellite link performance. The flare itself is a space weather event; its operational impact on terrestrial systems requires the composition.
co-06-aurora-forecast— Solar activity drives geomagnetic storms and aurorare-08-connection-quality— Solar storms can degrade satellite and HF radio communicationsk-01-anomaly-sentinel— Solar events can trigger automated monitoring alerts
Galaxies, nebulae, and star clusters visible tonight—with magnitude, angular size, altitude, and best viewing time. Uses a bundled Messier catalog with astronomy-engine for transit computation. The telescope owner's "what should I look at?" answer.
lat, lng (required); date (optional); limitingMag (optional, default 10.0)
Array of DSO results (name, type, magnitude, size, altitude, transit time), best targets subset
Bundled Messier catalog + astronomy-engine. Free. Status: dev.
Hourly. Object altitudes change through the night; catalog is static.
An observer with a 6" telescope near Bend, Oregon opens co-08 at 9pm in October. Best targets: M31 (Andromeda Galaxy, mag 3.4, transiting at 11pm at 78° altitude), M33 (Triangulum Galaxy, mag 5.7), M45 (Pleiades, rising in the east). The limiting magnitude of 10.0 for a 6" scope filters out objects too faint for the equipment. Stack with t-09-light-pollution (Bortle 3) and co-03-moon-phase (waning crescent rising at 3am—no interference tonight).
co-01-night-sky-map— Plots DSO positions on the star chart for visual orientationco-10-astrophotography-planner— DSO targets feed into the imaging session plant-09-light-pollution— Bortle class determines which DSOs are actually visible from the site
Visible satellite passes including Starlink trains, weather satellites, and GPS birds. Pass predictions with start/max/end positions, magnitude, and duration via N2YO. The moving points of light that aren't stars, aren't planes, and are increasingly numerous.
lat, lng (required); category (optional, satellite category filter)
Satellites currently above (position, altitude, velocity), upcoming pass predictions
N2YO API. Requires API key. Status: dev.
Every 6 hours. TLE orbital elements update periodically.
An astrophotographer setting up for a long exposure checks co-09 to plan around satellite trails. 14 satellites will cross the target field of view in the next 2 hours, including a Starlink train at 10:47pm. Knowing the timing allows pausing the capture during passes rather than discovering satellite streaks in post-processing. Stack with co-10-astrophotography-planner for the complete imaging session plan.
co-10-astrophotography-planner— Satellite timing avoids streak contamination in long exposuresco-05-iss-tracker— ISS is the brightest satellite; this rill covers everything elsea-08-astronomical-twilight— Satellites are visible only when sunlit in a dark sky (twilight hours)
The compound rill that ties the sky together: best imaging targets tonight with dark-sky windows, moon interference, Bortle class estimate, target ranking, and NASA's Astronomy Picture of the Day for inspiration. The session planner for anyone pointing a camera at the sky.
lat, lng (required); date (optional)
Ranked target list, dark window (start/end), Bortle class, APOD image data
Compound: astronomy-engine + NASA APOD (key required). Status: dev.
Hourly. Conditions change through the night; APOD updates daily.
A photographer plans a weekend trip to Oregon's Alvord Desert (Bortle 1—the darkest sky in the state). co-10 shows the dark window from 10:15pm to 4:45am, with moon below the horizon all night (new moon weekend). Top targets: Milky Way core (low in the south until midnight), M31 (high in the east all night), North America Nebula (overhead). The APOD features a stunning M31 image—a good omen. Stack with a-01-current-conditions and a-10-cloud-forecast for the go/no-go weather decision.
t-09-light-pollution— Bortle class sets the floor for imaging qualitya-10-cloud-forecast— Clear sky is the go/no-go condition for the entire sessiona-08-astronomical-twilight— Twilight timing defines the start of the dark windowco-03-moon-phase— Moon presence and illumination determine target selectionco-08-deep-sky-objects— DSO catalog provides the full target library to rank from
Biosphere — The Living World
Layered atop the physical substrate of Terra Firma is the living world—the plants that root in soil, the animals that move through habitat, the fungi that thread through forest floors, and the seasonal rhythms that govern them all. Biosphere rills transform coordinates into ecological portraits: what species have been observed here, what’s in bloom, what’s migrating through, what’s endangered, what’s edible, and what the land can sustain.
The data here is overwhelmingly community-sourced—iNaturalist’s 200M+ research-grade observations form the backbone of most rills, supplemented by government databases (EPA ecoregions, USDA hardiness zones, USA-NPN phenology) and specialized archives (eBird, OBIS, Xeno-canto, Movebank). These rills become most powerful in composition: a native plant atlas alone is a species list, but stacked with soil data, bloom calendars, and pollinator observations, it becomes a planting plan grounded in what actually grows and pollinates in this specific place.
flora (plant life, goddess of flowers)
What grows here, and what has always grown here? Flora answers the botanical questions about any location—from the USDA hardiness zone that determines which perennials survive winter, to the native species observed by community scientists, to the seasonal bloom rhythms that paint the calendar in flowers. It reaches into ecoregion classification, tree canopy change detection, fungal diversity, and pollinator garden planning. This is the family that transforms a pair of coordinates into a living botanical portrait of place.
USDA hardiness zones, native plant species identification, bloom seasonality, invasive species risk scoring, EPA ecoregion classification, tree canopy coverage and decadal change, phenological events (leaf-out, bloom, fruiting, fall color), fungal species and edibility, pollinator-plant relationships, and medicinal plant uses.
iNaturalist v1 API (species observations, free, 1 req/s). USDA PHZM API via phzmapi.org (hardiness zones, free). EPA Ecoregions MapServer (Level III/IV, free). MRLC GeoServer WMS (NLCD tree canopy, free). USA National Phenology Network (phenology, free). Nominatim (reverse geocoding, free). Xerces Society guidelines (curated pollinator-plant database, static).
A native plant list alone is a species inventory. Stack f-02-native-plant-atlas with t-02-soil-profile and f-01-hardiness-zone, and it becomes a planting recommendation grounded in what actually thrives in this soil and climate. Combine f-03-bloom-calendar with fa-05-pollinator-watch and f-09-pollinator-garden-planner, and you’ve composed a pollinator habitat assessment that reveals whether local bloom timing aligns with pollinator activity. Layer f-06-tree-canopy-coverage with f-04-invasive-species-alert, and canopy loss becomes a signal—was it development, drought, or an invasive pest wave?
Stack f-01-hardiness-zone + f-02-native-plant-atlas + f-03-bloom-calendar + f-09-pollinator-garden-planner for a property in Portland, Oregon. Zone 8b, with 247 native plant species observed within 25km. Bloom peaks in June but drops sharply after September—a forage gap for late-season pollinators. The pollinator planner recommends New England Aster and Canada Goldenrod to fill the August–October window, both confirmed locally by iNaturalist observations. This surfaces a signal that a garden designed for continuous bloom coverage would measurably support pollinator populations during the critical fall migration period.
| ID | Name | Description |
|---|---|---|
| f-01 | USDA Hardiness Zone | Plant hardiness zone, minimum winter temperature range, and zip code lookup |
| f-02 | Native Plant Atlas | Research-grade native plant observations ranked by frequency via iNaturalist |
| f-03 | Bloom Calendar | 12-month flowering timeline with relative bloom intensity per month |
| f-04 | Invasive Species Alert | Invasive plant detection with observation density and risk scoring |
| f-05 | Ecoregion Profile | EPA Level I–IV ecoregion classification for any US location |
| f-06 | Tree Canopy Coverage | NLCD tree canopy percentage with 2011–2021 change detection |
| f-07 | Phenology Tracker | Seasonal biological events—leaf-out, bloom, fruiting, fall color—via USA-NPN |
| f-08 | Mycology Finder | Fungi species near any location with edibility notes and seasonal fruiting patterns |
| f-09 | Pollinator Garden Planner | Native pollinator plant recommendations with bloom coverage analysis |
| f-10 | Medicinal Plant Guide | Traditional and medicinal plant uses by region (planned) |
Pinpoints the USDA Plant Hardiness Zone for any US location by reverse-geocoding coordinates to a zip code, then querying the PHZM database. The foundation rill for everything botanical—zones determine which perennials survive winter and feed directly into growing calendars, bloom predictions, and garden planning.
lat, lng (required)
Zone ID (e.g., “8b”), average annual minimum temperature range in °F, zip code used for lookup
USDA PHZM API via phzmapi.org (free). Nominatim reverse geocode (free). Status: stable.
24 hours. Hardiness zones are effectively static; caching is for API courtesy.
A gardener in Seattle’s Capitol Hill neighborhood queries f-01. Zone 8b, average annual minimum 15–20°F. This rules out tropical species but confirms that fig trees, rosemary, and lavender are viable perennials. Stack with cu-01-growing-calendar and the zone feeds directly into frost date calculations and planting windows—the difference between “start tomatoes indoors March 10” and “start tomatoes indoors April 20.”
f-02-native-plant-atlas— Zone sets the viability filter for which native species to recommendcu-01-growing-calendar— Zone determines frost dates that anchor the entire planting calendarf-09-pollinator-garden-planner— Zone filters the curated Xerces Society plant database to local candidates
Queries iNaturalist for research-grade plant observations near any location, returning a ranked list of native and naturalized species with photos, observation counts, and taxonomic family. The botanical census of what actually grows here—not what a field guide says should grow here, but what community scientists have photographed and confirmed.
lat, lng (required); radius in km (optional, default 25)
Species list: scientific name, common name, family, observation count, photo URL. Total distinct species count.
iNaturalist v1 API (free, 1 req/s rate limit). Research-grade observations only. Status: dev.
7 days. Plant observations accumulate slowly; weekly refresh captures seasonal additions.
A landscape designer surveys a 5-acre property outside Bend, Oregon. f-02 returns 183 native plant species within 25km, led by Ponderosa Pine (412 observations), Bitterbrush (287), and Oregon Grape (251). The top-10 list reveals a high-desert sagebrush steppe community—suggesting drought-adapted landscaping rather than the green lawns the client envisioned. Stack with t-02-soil-profile (sandy loam, well-drained) and the species list becomes a planting palette grounded in both ecology and soil chemistry.
f-05-ecoregion-profile— Ecoregion context explains why this specific species assemblage exists heret-02-soil-profile— Soil type constrains which observed species are actually plantable on this parcelf-04-invasive-species-alert— Separating native from invasive observations reveals ecological health
Builds a 12-month flowering timeline from iNaturalist observation density, showing when plants bloom near any location. Each month gets a relative intensity score (0–1) that reveals the seasonal rhythm of flowering—peak bloom in late spring, the midsummer lull, the fall aster surge.
lat, lng (required); radius in km (optional, default 25)
12 monthly entries with observation count and intensity (0–1). Peak bloom month with count. Total observations.
iNaturalist v1 histogram endpoint (free). Status: dev.
7 days. Bloom data is historical aggregate; new observations shift patterns slowly.
A beekeeper near Olympia, Washington checks f-03 to plan hive nutrition. Peak bloom: June (intensity 1.0), with a steep drop to October (0.15). The data suggests supplemental feeding may be needed from October through February. Stack with fa-05-pollinator-watch to correlate bloom timing with actual pollinator activity periods—if bees are still active in October but flowers aren’t, that’s a forage gap worth addressing.
fa-05-pollinator-watch— Compares bloom timing to pollinator activity; mismatches reveal forage gapsf-07-phenology-tracker— Phenology adds leaf-out, fruiting, and fall color to the bloom timelinefg-03-seasonal-harvest— Bloom correlates with harvest timing for fruit and berry foraging
Identifies invasive plant species near a location by querying iNaturalist for introduced species and cross-referencing against a curated list of 20 high-impact invasives. Risk is scored by observation density and species severity—from “low” (a few scattered observations) to “severe” (10+ invasive species or 500+ observations).
lat, lng (required); radius in km (optional, default 25)
Invasive species list with observation counts and severity. Risk level (low/moderate/high/severe). Total invasive observations and introduced species count.
iNaturalist v1 API with introduced=true filter (free). Cross-referenced against curated invasive species watchlist. Status: dev.
7 days. Invasive spread is gradual; weekly monitoring catches new establishment signals.
A property buyer evaluates a 20-acre parcel near Ashland, Oregon. f-04 returns risk level “high”—Scotch Broom (Cytisus scoparius, 187 observations) and Himalayan Blackberry (Rubus armeniacus, 342 observations) dominate. Both are aggressive colonizers of disturbed land. Combined with t-07-wildfire-history showing a 2020 burn perimeter overlapping the parcel, this surfaces a signal that post-fire invasive establishment may be underway—a factor in land management cost estimates.
t-07-wildfire-history— Post-fire landscapes are invasion hotspots; fire history predicts invasive riskf-06-tree-canopy-coverage— Canopy loss may signal invasive displacement of native forestcv-01-zoning-lookup— Some jurisdictions mandate invasive species removal on developed parcels
Identifies the EPA Level I through IV ecoregion for any US location by querying the EPA’s Ecoregions MapServer. Ecoregions classify land by shared ecological characteristics—the invisible boundaries that explain why the Willamette Valley grows different things than the Oregon High Desert 100 miles east.
lat, lng (required)
Four-level ecoregion hierarchy: Level I (continental), Level II (subcontinental), Level III (regional), Level IV (local). Each with code and name.
EPA Ecoregions MapServer (Level III and IV layer, free). US locations only. Status: dev.
365 days. Ecoregion boundaries are effectively static.
An ecologist preparing a site assessment near Hood River, Oregon queries f-05. Level III: “Cascades” (code 4). Level IV: “Western Cascades Lowlands and Valleys.” This context explains the species assemblage returned by f-02—the Douglas Fir and Western Red Cedar dominance is characteristic of this specific ecoregion, not a coincidence. Stack with t-02-soil-profile and the ecoregion adds the “why” to the soil’s “what.”
f-02-native-plant-atlas— Ecoregion explains why specific species dominate a locationt-02-soil-profile— Soil and ecoregion together define the growing environmentfa-04-endangered-species— Certain ecoregions harbor disproportionate endangered species concentrations
Samples NLCD tree canopy cover percentage at a point and computes the change between 2011 and 2021. Two raster queries, one number: how forested is this spot, and is it gaining or losing trees? The answer reshapes everything from urban heat island analysis to timber valuation.
lat, lng (required)
Canopy % (2021), canopy % (2011), percentage point change, direction (gain/loss/stable), human-readable interpretation.
MRLC GeoServer WMS (NLCD Tree Canopy Cover, free). US locations only. Status: dev.
365 days. NLCD datasets update roughly every 3–5 years; the 2021 vintage is current.
A climate researcher checks f-06 across a grid of points in the Portland metro area. Downtown: 12% canopy (2021), down from 18% (2011)—a 6-point loss. Forest Park: 89%, stable. The pattern surfaces an urban canopy equity question: neighborhoods losing tree cover may experience higher summer temperatures. Stack with t-11-air-quality and vl-01-property-valuation and the canopy data becomes a factor in both environmental justice analysis and property value modeling.
t-11-air-quality— Tree canopy improves air quality; loss may correlate with AQI degradationf-04-invasive-species-alert— Canopy loss + invasive presence may signal ecological displacementt-01-helio-study— Canopy percentage modifies effective solar exposure on a parcel
Tracks the five seasonal milestones—leaf-out, open flowers, fruiting, colored leaves, and leaf drop—using six years of observation data from the USA National Phenology Network. Phenology is the science of seasonal timing, and it turns out the calendar is moving: spring arrives earlier, fall lingers longer.
lat, lng (required); radius in degrees (optional, default 0.5° ≈ 55km)
Phenophase groups with monthly distribution and species counts. Current active season. Total species tracked. Observation period (2020–2025).
USA National Phenology Network (getSummarizedData, free). Status: dev.
7 days. Phenology data is observational; new records trickle in throughout the season.
A landscape photographer planning a fall color trip checks f-07 for the Columbia River Gorge. Colored leaves peak in the third week of October, with 67% of observations between Oct 10–Nov 5. Leaf drop follows two weeks later. Combined with a-10-cloud-forecast for clear-sky probability and t-03-elevation-slope for south-facing viewpoints, this produces a precision timing window for peak fall color photography in the Gorge.
f-03-bloom-calendar— Bloom calendar shows when flowers appear; phenology adds the full seasonal arca-01-current-conditions— Weather affects phenological timing; warm springs shift leaf-out earliercu-01-growing-calendar— Phenological signals like last frost refine planting date windows
Discovers mushroom and fungi species near any location using iNaturalist research-grade observations. Returns species photos, observation counts, seasonal fruiting patterns, and edibility notes from a built-in genus reference with safety warnings. The bridge between ecology and foraging—though the edibility notes come with big caveats.
lat, lng (required); radius in km (optional, default 25); month (optional, 1–12)
Fungi species with scientific/common names, observation counts, photos, and edibility notes (edible/not edible/unknown + caution text). Peak fruiting month. 12-month seasonality histogram.
iNaturalist v1 API (free). Edibility reference: curated 10-genus lookup (Cantharellus, Boletus, Amanita, Galerina, Agaricus, Morchella, Pleurotus, Trametes, Laetiporus, Coprinus). Status: dev.
7 days. Fungi observations are seasonal; weekly refresh tracks fruiting waves.
A forager checks f-08 near Corvallis, Oregon in November. Top species: Cantharellus formosus (Pacific Golden Chanterelle, 89 observations), Trametes versicolor (Turkey Tail, 67), and Amanita muscaria (Fly Agaric, 41). The edibility reference flags Chanterelle as edible (“confirm ID carefully—Jack O’Lantern look-alikes exist”) and Amanita as not edible. Stack with fg-01-mushroom-finder for the full foraging safety layer with toxic look-alike warnings.
fg-01-mushroom-finder— Mushroom Finder adds PFAF edibility ratings and toxic look-alike cross-referencingfg-09-weather-window— Mushroom fruiting requires specific humidity and temperature conditionst-02-soil-profile— Soil type influences which mycorrhizal species occur
Recommends native plants that support local pollinators by cross-referencing iNaturalist observations with a curated 20-species database built from Xerces Society guidelines. Calculates 12-month bloom coverage and flags forage gaps where additional species are needed. The compound rill that turns ecology into a planting checklist.
lat, lng (required); radius in km (optional, default 25)
Recommended plants sorted by bloom start (with local observation confirmation). 12-month bloom coverage (plant count per month). Forage gap months. Unique pollinator species supported. Total recommended count.
iNaturalist v1 API (free) + Xerces Society curated pollinator-plant database (20 species, static). Status: dev.
7 days. Observation data is the dynamic layer; the curated database is static.
A homeowner in Eugene, Oregon wants to replace a lawn with a pollinator garden. f-09 recommends 16 of 20 species as locally viable, with 12 confirmed by iNaturalist observations. Bloom coverage shows April–September well-covered, but October has zero blooming species—a forage gap during Monarch butterfly fall migration. Adding Azure Blue Sage and Canada Goldenrod fills the gap. The planner supports 14 unique pollinator species including Monarch butterfly, bumblebees, and hummingbirds.
fa-05-pollinator-watch— Confirms which pollinators are actually active locally to match plants to visitorsf-01-hardiness-zone— Zone determines which recommended plants survive winter in this locationt-02-soil-profile— Soil chemistry determines which recommended plants will actually thrivef-03-bloom-calendar— Bloom calendar validates the coverage analysis against real observation data
Traditional and medicinal plant uses by region, including preparation methods, phytochemical data, and safety notes. Architecturally planned but not yet implemented—the directory structure exists, but manifest and API files have not been created.
lat, lng (required); species (optional name lookup)
Planned: plant list with medicinal ratings, traditional uses, preparation methods, phytochemical data, safety warnings.
Planned: iNaturalist (occurrence) + curated ethnobotanical database. Not yet implemented.
Planned: 30 days for occurrence data; curated data is static.
Planned: a herbalist checks which medicinal plants grow near a rural property, with traditional preparation methods and modern safety contraindications. The Forage family’s fg-08-medicinal-foraging rill covers similar ground with a foraging-safety lens and is fully implemented—this rill will extend that work with a Flora-family ecological perspective.
fg-08-medicinal-foraging— Forage family covers medicinal foraging with safety layer; this rill adds ecological contextf-02-native-plant-atlas— Native plant observations provide the occurrence foundationf-05-ecoregion-profile— Ecoregion determines which medicinal traditions are geographically relevant
fauna (animal life, woodland creatures)
What moves through this place? Fauna answers the animal questions—from the birds singing at dawn, to the mammals crossing trails at dusk, to the pollinators visiting flowers at midday, to the marine life cruising offshore. It spans the full spectrum of animal observation: real-time eBird sightings, iNaturalist community science records, Movebank GPS migration tracks, IUCN Red List conservation assessments, Xeno-canto bird recordings, and OBIS ocean biodiversity data. This is the family that populates a landscape with its inhabitants.
Bird species sightings and seasonal presence, mammal and reptile observations, GPS-tracked migration routes, endangered species with IUCN/ESA status, pollinator diversity and seasonal activity, marine and coastal species occurrence, bird vocalizations and sonograms, trail camera species predictions, insect and arachnid diversity, and animal track identification.
iNaturalist v1 API (wildlife observations, free). eBird API v2 (bird sightings, key required). IUCN Red List API v4 (conservation status, key required). OBIS v3 (marine biodiversity, free). Xeno-canto API v3 (bird recordings, free). Movebank API (GPS migration tracks, auth required). USFWS ECOS (US endangered species, free). Curated track/sign database (static).
A bird species list alone is a checklist. Stack fa-01-bird-watcher with f-06-tree-canopy-coverage and t-09-light-pollution, and it becomes a habitat quality assessment—declining bird diversity may signal canopy loss or light encroachment. Combine fa-04-endangered-species with cv-01-zoning-lookup and a property evaluation gains a regulatory dimension: ESA-listed species within range can trigger development review requirements. Layer fa-05-pollinator-watch with f-09-pollinator-garden-planner and the composition matches actual pollinator populations to recommended garden plants.
Stack fa-01-bird-watcher + fa-02-wildlife-atlas + fa-04-endangered-species + f-05-ecoregion-profile for a proposed development site near the Sandy River, Oregon. eBird shows 47 species in the last 30 days including Great Blue Heron and Bald Eagle. Wildlife Atlas returns 23 mammal species. Endangered Species Alert flags one IUCN-Vulnerable species: Western Pond Turtle. Ecoregion: Willamette Valley / Puget Trough. This surfaces a signal that the riparian corridor supports significant biodiversity, and the turtle presence may warrant a Section 7 consultation before permitting—a finding invisible without the cross-family stack.
| ID | Name | Description |
|---|---|---|
| fa-01 | Bird Watcher | Recent bird sightings near any location with species, counts, and notable observations via eBird |
| fa-02 | Wildlife Atlas | Mammal, reptile, and amphibian species with photos, observation counts, and conservation status |
| fa-03 | Migration Tracker | GPS migration tracks and seasonal presence data from Movebank and eBird |
| fa-04 | Endangered Species Alert | Threatened and endangered species with IUCN, ESA, and NatureServe conservation ranks |
| fa-05 | Pollinator Watch | Bee, butterfly, and moth diversity with seasonal activity histograms |
| fa-06 | Marine Life Guide | Coastal and marine species from OBIS ocean biodiversity records |
| fa-07 | Bird Song ID | Bird call recordings, sonograms, and species identification via Xeno-canto |
| fa-08 | Wildlife Camera Log | Expected species for trail camera locations with activity period predictions |
| fa-09 | Insect Atlas | Local insect and arachnid diversity with seasonal activity patterns |
| fa-10 | Animal Track Guide | Track dimensions, gait patterns, scat descriptions, and habitat sign for North American mammals |
Recent bird sightings near any location via eBird, the world’s largest citizen science bird database. Returns species lists with observation counts, notable sightings, and seasonal presence derived from 30-day observation windows. The starting point for any avian inquiry.
lat, lng (required); radius (optional); days (optional, observation window)
Species list with common/scientific names, observation counts, and notable sightings. Total species count.
eBird API v2 (Cornell Lab of Ornithology, key required). Status: dev.
4 hours. Bird observations are submitted in near-real-time; frequent refresh captures daily activity.
A birder visiting Ridgefield National Wildlife Refuge near Portland checks fa-01. 52 species reported in the last 30 days, led by Cackling Goose (2,400+ observations), Northern Pintail, and Bald Eagle. The notable sightings flag a Tundra Swan—uncommon here. Stack with fa-03-migration-tracker and the sightings gain temporal context: the Cackling Geese are wintering, not resident, and will depart by March.
fa-03-migration-tracker— Migration data adds temporal context: which sightings are resident vs. migratoryfa-07-bird-song-id— After identifying a species by sight, the recording library helps learn its callf-06-tree-canopy-coverage— Bird diversity correlates with canopy coverage; loss may explain declining counts
Mammals, reptiles, and amphibians observed near any location, pulled from iNaturalist’s research-grade observations. Each species includes photos, observation counts, conservation status, and taxonomic classification. The non-avian counterpart to Bird Watcher—everything with fur, scales, or moist skin.
lat, lng (required); radius in km (optional, default 25); month (optional, 1–12)
Species list with photos, observation counts, conservation status, and taxonomic class. Taxa breakdown (Mammalia/Reptilia/Amphibia counts).
iNaturalist v1 API (free). Filters: Mammalia, Reptilia, Amphibia. Research-grade only. Status: dev.
4 hours. Wildlife activity varies by time of day and season.
An environmental consultant surveys a proposed solar farm site near Prineville, Oregon. fa-02 returns 31 wildlife species: 18 mammals (Mule Deer, Pronghorn, Black Bear), 9 reptiles (Western Fence Lizard, Gopher Snake), and 4 amphibians (Pacific Treefrog, Western Toad). The Western Toad observation is significant—stack with fa-04-endangered-species to check its conservation status, which may affect permitting.
fa-04-endangered-species— Conservation status layer identifies which observed species carry regulatory weightfa-08-wildlife-camera— Observed species feed trail camera placement and timing optimizationfa-10-animal-track— Track guide covers the mammals identified here with field sign details
GPS migration tracks and seasonal presence data by combining Movebank satellite tracking with eBird observation records. Shows actual migration routes with timestamped waypoints, arrival/departure windows, and whether a species is resident, breeding, wintering, or passing through.
lat, lng (required); species (optional); season (optional: spring, fall, all)
GPS tracks with waypoints and total distance. Seasonal presence classifications (resident/breeding/winter/migrant). Nearby Movebank tracking studies with accessibility status.
Movebank REST API (auth required, 1 concurrent req/IP). eBird API v2 (key required). Status: dev.
Daily. Migration timing shifts year-to-year; daily refresh catches arrival/departure signals.
A wildlife biologist checks fa-03 for the Willamette Valley in March. eBird seasonal data classifies Cackling Goose as “winter” (present Nov–Feb, departing), Barn Swallow as “breeding” (arriving), and Dark-eyed Junco as “resident” (year-round). Movebank shows 3 public GPS studies within 200km. The compound picture reveals the valley as a continental-scale intersection: Arctic-breeding geese leaving as Central American swallows arrive—a migration crossroads visible only through data fusion.
fa-01-bird-watcher— Recent sightings gain temporal meaning when overlaid on migration patternsa-01-current-conditions— Weather systems trigger migration waves; storms may concentrate or delay arrivalst-09-light-pollution— Nocturnal migrants are disoriented by light pollution; dark corridors matter
Layered conservation intelligence from three authoritative sources: IUCN Red List (global), USFWS ECOS (US federal ESA), and NatureServe (state-level ranks). Fetches assessments in parallel and merges them into a unified view sorted by severity. The rill that determines whether a location carries regulatory conservation weight.
lat, lng (required); country (optional, ISO alpha-2, default derived from coordinates)
Species list with unified conservation status (IUCN category, ESA status, NatureServe rank, population trend, severity color). Breakdown by IUCN category. Total endangered count.
IUCN Red List API v4 (key required). USFWS ECOS (US listings, free). Nominatim (country resolution, free). Status: dev.
90 days. IUCN assessments update annually at most; ESA listings change on regulatory timelines.
A developer evaluating a coastal site near Bandon, Oregon runs fa-04. The alert returns 8 species: Western Snowy Plover (IUCN Near Threatened + ESA Threatened), Marbled Murrelet (IUCN Endangered + ESA Threatened), and 6 others. The Plover listing is critical—nesting habitat on beaches can trigger seasonal construction restrictions and buffer zones. Combined with cv-01-zoning-lookup, the endangered species data transforms a property assessment from “buildable” to “buildable with significant regulatory constraints.”
cv-01-zoning-lookup— ESA-listed species can trigger Section 7 consultation requirements for developmentfa-02-wildlife-atlas— Wildlife observations identify which species are actually present, not just listedf-05-ecoregion-profile— Ecoregion context determines which endangered species are expected in an area
Bee, butterfly, and moth species near any location with 12-month activity histograms. Queries iNaturalist for Lepidoptera and Hymenoptera in parallel, merging the results into a unified pollinator diversity profile. Peak months are flagged automatically when activity exceeds 70% of maximum.
lat, lng (required); radius in km (optional, default 25); month (optional, 1–12)
Pollinator species with photos, counts, and order (Lepidoptera/Hymenoptera). 12-month activity histogram with peak months. Order breakdown.
iNaturalist v1 API (free). Taxon filters: Lepidoptera (47157) and Hymenoptera (47201). Status: dev.
4 hours. Pollinator observations are highly seasonal and weather-dependent.
An organic farmer near Salem, Oregon checks fa-05 in July. 67 pollinator species observed: 41 Lepidoptera (Western Tiger Swallowtail, Painted Lady, various skippers) and 26 Hymenoptera (bumblebees, mason bees, paper wasps). Peak months: June–August. Stack with f-09-pollinator-garden-planner to determine which plants would sustain this pollinator community beyond August—extending the bloom season directly supports the farm’s pollination needs.
f-09-pollinator-garden-planner— Matches actual pollinator populations to recommended garden plantsf-03-bloom-calendar— Compares pollinator activity timing to flower availabilitycu-03-companion-planting— Pollinator-attracting companions can increase crop yields
Coastal and marine species from OBIS, the Ocean Biodiversity Information System, supplemented with iNaturalist photos for common species. Queries a WKT bounding box around any coastal location against 100M+ ocean occurrence records spanning 160,000+ marine species. The underwater counterpart to the terrestrial wildlife rills.
lat, lng (required); radius in km (optional, default 50)
Marine species with scientific/common names, occurrence counts, and photos. Family breakdown. Total species and occurrences. Search area description.
OBIS v3 API (free). iNaturalist v1 (Actinopterygii + Mollusca photos, free). Status: dev.
7 days. OBIS aggregates historical records; new data arrives in batches.
A marine educator prepares a tide pool field trip near Depoe Bay, Oregon. fa-06 returns 142 marine species within 50km, including Purple Sea Urchin, Ochre Sea Star, and Giant Pacific Octopus. Family breakdown shows Gastropoda (23 species) and Actinopterygii (31 species) dominating. Stack with h-02-tidal-data for low tide windows—the best tide pool visibility occurs at minus tides, and knowing which species to look for transforms the visit from wandering to targeted discovery.
h-02-tidal-data— Tide windows determine when marine species are accessible in intertidal zonesfg-06-seaweed-coastal— Seaweed and marine life co-occur; the composition builds a complete coastal profilefa-04-endangered-species— Marine species may carry IUCN listings affecting coastal development
Bird call recordings, sonograms, and playback from Xeno-canto, the world’s largest open-access bird sound library with 900,000+ recordings across 10,800+ species. Searches by location, species, vocalization type (song, call, alarm), and quality rating. The auditory layer of bird identification.
lat, lng (required); species (optional); type (optional: song, call, alarm call, flight call)
Recordings with audio URL, sonogram URL, quality rating (A–E), duration, recordist credit, and location. Top species by recording count.
Xeno-canto API v3 (free, CC-licensed recordings). Status: dev.
30 days. The recording archive grows but doesn’t change retroactively.
A birder hears an unfamiliar call on an early morning hike near Mount Tabor in Portland. They query fa-07 for the area with quality filter “A.” The top results include Varied Thrush (song), Steller’s Jay (call), and Pacific Wren (song). Playing the recordings against memory narrows the ID to Pacific Wren—a remarkably complex song from a tiny bird. Stack with fa-01-bird-watcher to confirm the species has been reported in this area recently.
fa-01-bird-watcher— Visual sightings complement audio identification for species confirmationsn-05-natural-soundscape— Bird songs are a primary component of ambient natural soundscapest-10-noise-profile— Anthropogenic noise masks bird calls; noise data predicts audibility
Predicts which animals to expect at a trail camera location based on local iNaturalist mammal and bird observations. Classifies species by activity period (nocturnal, crepuscular, diurnal) using taxonomic family inference, and generates camera setup advice—including whether to enable infrared mode for the predominantly nocturnal cast.
lat, lng (required); radius in km (optional, default 15); month (optional, 1–12)
Expected species with activity periods and photos. 12-month activity histogram. Camera setup advice text. Taxa breakdown.
iNaturalist v1 API (free). Activity period inferred from taxonomic family (curated). Status: dev.
4 hours. Wildlife observations are seasonal; refresh captures recent activity.
A landowner places a trail camera on a 40-acre property near Estacada, Oregon. fa-08 returns 24 expected species: Black Bear (crepuscular), Raccoon (nocturnal), Black-tailed Deer (crepuscular), Coyote (crepuscular), and Barred Owl (nocturnal) lead the list. Camera advice: “Enable infrared mode—most local species are nocturnal or crepuscular.” Peak activity: May–October. Stack with fa-10-animal-track to learn the field sign that confirms camera subjects without reviewing footage.
fa-10-animal-track— Track identification confirms camera species and reveals animals that avoid camerasfa-02-wildlife-atlas— Broader species data feeds into the camera prediction modelt-09-light-pollution— Dark sites have more nocturnal wildlife activity; Bortle class predicts camera yield
Local insect and arachnid diversity from iNaturalist, covering beetles, butterflies, moths, flies, ants, spiders, and everything else with an exoskeleton. Returns species with order and family classification, seasonal activity histograms, and diversity breakdowns. The overlooked majority of animal life, finally counted.
lat, lng (required); radius in km (optional, default 25); month (optional, 1–12)
Insect/arachnid species with photos, counts, order, and family. Seasonal activity histogram. Order breakdown (Coleoptera, Lepidoptera, Diptera, etc.). Class breakdown (Insecta vs. Arachnida).
iNaturalist v1 API (free). Filters: Insecta and Arachnida. Research-grade only. Status: dev.
4 hours. Insect activity is temperature-sensitive and changes throughout the day.
An entomologist surveying a meadow restoration project near Corvallis, Oregon queries fa-09. 178 arthropod species observed: Coleoptera (41 species), Lepidoptera (37), Hymenoptera (29), Diptera (24), Araneae (19). Peak activity: June–August. The species richness, especially in beetles and butterflies, suggests the meadow restoration is succeeding—insect diversity is a sensitive indicator of habitat health. Stack with fa-05-pollinator-watch to separate pollinator species from the broader insect community.
fa-05-pollinator-watch— Overlapping taxa; Insect Atlas provides the full picture, Pollinator Watch focuses on ecosystem servicescu-06-pest-disease-identifier— Some observed insects are crop pests; the atlas provides ecological contextf-06-tree-canopy-coverage— Canopy cover and insect diversity are positively correlated in most ecosystems
A curated field guide to animal tracks and sign for 12 common North American mammals, filtered to species confirmed in your area by iNaturalist. Each entry includes front and hind track dimensions, gait pattern, stride length, scat description, behavioral sign, habitat preferences, and seasonal notes. The field guide in your pocket, localized.
lat, lng (required); species (optional, name filter)
Track profiles: front/hind track dimensions, toe count, claw visibility, gait pattern, stride, scat shape and contents, habitat sign. Regional presence count (how many guide species are confirmed locally).
Curated static track database (12 species). iNaturalist (regional presence confirmation, free). Status: dev.
30 days. Curated data is static; iNaturalist presence confirmation refreshes monthly.
A hiker finds unfamiliar tracks on a muddy trail near Silver Falls State Park, Oregon. fa-10 returns 10 of 12 guide species as regionally present. The track is round, about 2” wide, with no claw marks—matching Bobcat (retractable claws, round print, 4 toes). Stride is 14”, consistent with the diagonal-walk gait pattern. Nearby scratch marks on soil confirm territorial marking behavior. Stack with fa-02-wildlife-atlas—Bobcat has 23 iNaturalist observations within 25km, confirming established presence in this area.
fa-02-wildlife-atlas— Atlas confirms species presence; tracks provide evidence of animals that avoid camerasfa-08-wildlife-camera— Track identification guides camera placement to confirmed animal travel routesct-01-terrain-map— Topography and water features predict where animals travel and where tracks concentrate
forage (wild food, gathered sustenance)
What can you eat here? Forage is the wild edible intelligence layer—a three-tier system combining an occurrence layer (where species grow, via iNaturalist and GBIF), an edibility layer (is it safe to eat, via PFAF and Trefle), and a safety layer (toxic look-alikes, confidence scoring, and mandatory disclaimers). This family handles the highest-stakes data in the entire Biosphere domain: misidentification of toxic species has resulted in documented hospitalizations and deaths. Every rill carries mandatory safety disclaimers and confidence scoring as architectural requirements, not optional features. The root safety rill, fg-04-toxicity-checker, feeds every other rill in the family.
Edible mushroom species by location and season, wild edible plants with edibility ratings, seasonal harvest calendars, toxic look-alike warnings with distinguishing features, urban fruit and nut tree maps, edible seaweed and coastal species, nut and berry ripeness tracking, medicinal plant properties and phytochemistry, weather-optimized foraging windows, and preservation methods (drying, tincture, fermentation).
iNaturalist v1 API (occurrence data, free). GBIF species match (taxonomic resolution, free). Trefle API (botanical data, token required). PFAF (Plants For A Future, static curated). Dr. Duke’s Phytochemical DB (ethnobotanical data, static). Mushroom Observer (expert ID, free). Falling Fruit API (urban foraging map, key required). Open-Meteo (weather forecasts, free). Curated static datasets for toxic look-alikes (21 pairs), seaweed species (15), and preservation methods (9 entries).
A mushroom species list alone is dangerous—Chanterelles and Jack-o’-Lanterns look alike. Stack fg-01-mushroom-finder with fg-04-toxicity-checker and every result carries a toxic look-alike warning and confidence score. Layer fg-09-weather-window with fg-01 and you get optimal foraging conditions: mushrooms fruit 2–5 days after rain at the right temperature. Combine fg-02-wild-edible-atlas with fg-03-seasonal-harvest and a walk through a meadow becomes a seasonal harvest plan. Stack fg-06-seaweed-coastal with a-03-tide-tables and coastal foraging gains tidal safety context. This family always composes with safety layers first, adventure second.
Stack fg-01-mushroom-finder + fg-04-toxicity-checker + fg-09-weather-window + fg-10-preservation-guide for a chanterelle expedition near Tillamook, Oregon. Mushroom Finder returns 14 fungi species within 50km, led by Golden Chanterelle (89 observations) and King Bolete (34). Toxicity Checker flags the Chanterelle/Jack-o’-Lantern pair: “Jack-o’-Lantern gills are non-forked, bioluminescent; Chanterelle gills are forked, non-luminous.” Weather Window scores tomorrow as GO (82/100): 15°C, 85% humidity, 3 rain days in the last week. Preservation Guide provides dehydrator drying instructions (5 steps, 1-year shelf life). The composition transforms a hike into a planned, safety-screened foraging operation.
| ID | Name | Description |
|---|---|---|
| fg-01 | Mushroom Finder | Edible mushroom species by location and season with toxic look-alike warnings and safety disclaimers |
| fg-02 | Wild Edible Atlas | Wild edible plants by region with edibility ratings, hazard warnings, and photo IDs |
| fg-03 | Seasonal Harvest Calendar | 12-month foraging timeline with species peak windows and intensity scores |
| fg-04 | Toxicity Checker | Toxic look-alike warnings, compound data, and confidence scoring — the root safety rill |
| fg-05 | Urban Foraging Map | Mapped fruit trees, nut trees, and edible plants in urban areas via Falling Fruit |
| fg-06 | Seaweed & Coastal Foraging | Edible seaweed and coastal species with harvest conditions and tidal context |
| fg-07 | Nut & Berry Tracker | Wild nut and berry species with ripeness timing and seasonal availability |
| fg-08 | Medicinal Foraging Guide | Wild medicinal plants with phytochemistry, traditional uses, and preparation methods |
| fg-09 | Forager’s Weather Window | Optimal weather conditions for foraging targets with go/no-go scoring |
| fg-10 | Preservation Guide | Drying, tincture, fermentation, and smoking methods with step-by-step instructions |
Finds edible mushroom species near a location by querying iNaturalist for Fungi occurrences and cross-referencing a curated 16-species edibility database covering both choice edibles (Chanterelle, Morel, King Bolete) and deadly species (Death Cap, Destroying Angel). Every result passes through fg-04’s toxic look-alike matcher and confidence scorer. The highest-risk rill in the Forage family—mushroom misidentification is the leading cause of foraging fatalities.
lat, lng (required); radius in km (optional, default 50); month (optional, 1–12)
Mushroom species list with common/scientific names, edibility ratings (1–5), observation counts, photo URLs, toxic look-alike warnings, and mandatory safety disclaimer with Poison Control number.
iNaturalist v1 species_counts endpoint (Fungi, free, 1 req/s). Mushroom Observer (expert IDs). PFAF edibility cross-reference (static curated). 7-day cache. Status: dev.
7 days. Fungal fruiting is seasonal and weather-dependent; weekly refresh captures fruiting flushes.
A mycologist plans an October foray in the Tillamook State Forest, Oregon. fg-01 returns 14 species within 50km, led by Cantharellus formosus (Pacific Golden Chanterelle, 89 observations, edibility: 5) and Boletus edulis (King Bolete, 34 observations). The look-alike matcher flags: “Chanterelle → Jack-o’-Lantern (Omphalotus olearius): forked gills vs. non-forked, no bioluminescence vs. bioluminescent.” Stack with fg-09-weather-window—score 82/100 after three days of rain—and the foray has both species targets and optimal timing.
fg-04-toxicity-checker— Every mushroom result passes through the toxicity pipeline for look-alike warnings and confidence scoringfg-09-weather-window— Mushrooms fruit 2–5 days after rain; weather data transforms a species list into a timing recommendationfg-03-seasonal-harvest— Histogram data reveals which months each species fruits most heavilyfg-10-preservation-guide— Dehydrator drying and sauté-freeze methods for preserving a large haul
Queries iNaturalist for Plantae occurrences near a location and cross-references a curated 15-species edibility database covering common wild edibles (Stinging Nettle, Dandelion, Chickweed, Burdock) through to species requiring extreme caution (Wild Carrot, Ground Elder). Only species with verified edibility data are returned—unknowns are suppressed, not guessed. Two species carry EXTREME CAUTION flags for deadly look-alikes.
lat, lng (required); radius in km (optional, default 50); edibility_min (optional, default 3, scale 1–5)
Wild edible species list with edibility rating, hazard warnings, observation counts, photos. Only curated matches returned. Safety disclaimer.
iNaturalist v1 species_counts (Plantae, free). PFAF edibility database (static). NatureServe (conservation status). 7-day cache. Status: dev.
7 days. Plant observations accumulate slowly; weekly refresh balances freshness with API courtesy.
A foraging instructor prepares a class walk through Forest Park, Portland. fg-02 returns 8 edible species within 25km above the edibility threshold of 3: Stinging Nettle (Urtica dioica, edibility 5), Dandelion (edibility 5), and Chickweed (edibility 4) lead the list. Wild Carrot (Daucus carota) triggers an EXTREME CAUTION flag: “Deadly look-alike Poison Hemlock (Conium maculatum).” The instructor uses this to structure the walk: safe species first, then a teaching moment on hemlock identification. Stack with fg-04-toxicity-checker to get full distinguishing-feature descriptions.
fg-04-toxicity-checker— Detailed look-alike warnings with distinguishing features for every flagged speciesfg-03-seasonal-harvest— Seasonality data shows when each edible plant is at its peakfg-07-nut-berry-tracker— Nut and berry species extend the edible plant list with fruiting-body datafg-08-medicinal-foraging— Many wild edibles double as medicinal plants; medicinal data adds a use dimension
Builds a 12-month foraging timeline from iNaturalist observation histograms, tracking 12 key species (Chanterelle, Morel, Blackberry, Elderberry, Stinging Nettle, and others) to reveal when each peaks near a given location. Peak months are defined as months reaching ≥50% of maximum observation count. The seasonal rhythm of wild food availability, derived from actual observation data rather than textbook generalizations.
lat, lng (required); radius in km (optional, default 50); month (optional, defaults to current)
12 monthly summaries with species counts and normalized intensity (0–1). In-season-now species list. Per-species peak months.
iNaturalist v1 histogram endpoint with interval=month_of_year (free, sequential fetches respecting 1 req/s). Trefle (supplemental). 7-day cache. Status: dev.
7 days. Histogram data is a historical aggregate that shifts slowly as new observations accumulate.
A wild food enthusiast near Eugene, Oregon checks fg-03 in October. Three species are in season now: Golden Chanterelle (peak Sep–Nov, intensity 0.9), Oregon Grape (peak Aug–Oct, intensity 0.6), and Chicken of the Woods (peak Sep–Oct, intensity 0.5). The calendar shows a lean period from January through March with near-zero intensity. Stack with fg-09-weather-window to narrow October down to the best day this week for chanterelles (2–5 days after sustained rain, 10–18°C).
fg-01-mushroom-finder— Mushroom occurrence data feeds the fungal portion of the seasonal calendarfg-02-wild-edible-atlas— Plant occurrence data feeds the wild edible portionfg-09-weather-window— Weather scoring narrows “in season” to “in season and conditions are right today”f-03-bloom-calendar— Bloom timing often correlates with fruit and berry harvest timing
The root safety rill for the entire Forage family. Resolves a species name to canonical taxonomy via GBIF, cross-references PFAF edibility data (15 species), Trefle botanical data, and a curated database of 21 toxic look-alike pairs (10 fungi, 8 plants, 2 berries, 1 coastal). Exports shared safety utilities consumed by every other forage rill: matchToxicLookalike(), calculateConfidence(), shouldSuppressEdibility(), and buildSafetyDisclaimer(). Includes US Poison Control number (1-800-222-1222) in every output.
species (required, common or scientific name)
Toxicity level (safe/caution/toxic/deadly), toxic compounds list, look-alike warnings with distinguishing features, edibility confidence (high/medium/low/insufficient), conservation status, and mandatory safety disclaimer.
GBIF species match (taxonomy, free). Trefle API (botanical data, token required). PFAF (edibility, static curated). NatureServe (conservation, free). 30-day cache. Status: dev.
30 days. Toxicological data changes infrequently; long cache is appropriate for safety-verified content.
A forager finds what they think is Wild Garlic (Allium ursinum). fg-04 returns toxicity: “safe” but flags TWO deadly look-alikes: Lily of the Valley (Convallaria majalis, cardiac glycosides) and Autumn Crocus (Colchicum autumnale, colchicine). Distinguishing features: “Wild Garlic has a strong garlic scent when crushed; Lily of the Valley is odorless; Autumn Crocus has no leaf stalk.” Confidence: MEDIUM (2 sources, look-alikes present). The safety disclaimer reads: “Never consume wild plants based solely on digital identification. Always confirm with an experienced forager or mycologist.”
fg-01-mushroom-finder— Every mushroom result passes through toxicity checking before displayfg-02-wild-edible-atlas— Every plant result passes through toxicity checking before displayfg-06-seaweed-coastal— Coastal species pass through the same safety pipelinefg-08-medicinal-foraging— Medicinal plants checked for toxic compounds and drug interactions
Surfaces mapped fruit trees, nut trees, and edible plants in urban areas using the Falling Fruit community database. Returns foraging spots with species type, GPS coordinates, distance from the user, and community-contributed access notes. Includes a curated 10-spot sample dataset (San Francisco Bay Area) as fallback when API credentials are absent.
lat, lng (required); radius in km (optional, default 5); type (optional filter: fruit/nut/herb)
Foraging spot list with species, coordinates, distance, access notes, community reviews. Sorted by proximity.
Falling Fruit API v0.3 (community-mapped edible plants, key required). Bounding-box search. 30-day cache. Status: dev.
30 days. Urban tree locations change slowly; community additions are the primary update vector.
A neighborhood food-security researcher maps edible resources in Southeast Portland. fg-05 returns 23 spots within 5km: fig trees (7), apple trees (6), plum trees (4), hazelnut trees (3), rosemary hedges (3). Each spot includes GPS coordinates and access notes (“street tree, public sidewalk” vs. “backyard, ask first”). Stack with fg-03-seasonal-harvest to determine which trees are producing now and which are months away from fruiting.
fg-04-toxicity-checker— Verifies that mapped species are genuinely edible, not misidentifiedfg-03-seasonal-harvest— Seasonal data reveals which mapped trees are actually producing right nowfg-07-nut-berry-tracker— Nut and berry ripeness timing for mapped hazelnut and berry bushes
Identifies edible seaweed and coastal plant species near any coastline using a curated database of 15 species (10 seaweeds, 5 coastal plants) enriched with iNaturalist occurrence data. Covers Sea Lettuce, Dulse, Nori, Sugar Kelp, Bladderwrack, Wakame, and coastal plants like Samphire and Sea Beet. Confidence scoring reflects whether species have been observed nearby (MEDIUM) or only exist in the curated database (LOW).
lat, lng (required); radius in km (optional, default 50); month (optional)
Coastal species list with edibility rating, harvest season, preparation notes, iNaturalist observation count, confidence level. Safety disclaimer.
iNaturalist v1 (Chromista + Plantae, free). Curated seaweed database (15 species, static). GBIF (taxonomy). 7-day cache. Status: dev.
7 days. Coastal species observations are sparse; weekly refresh captures new community science records.
A chef researching local coastal ingredients near Cannon Beach, Oregon checks fg-06. Results: Bull Kelp (Nereocystis luetkeana, 23 iNaturalist observations, edibility 4, harvest season Jul–Oct), Sea Lettuce (Ulva lactuca, 15 observations, edibility 4, year-round), and Samphire (Salicornia europaea, 8 observations, edibility 5, Jun–Sep). The toxic look-alike for Sea Lettuce—Acid Kelp—is flagged via fg-04. Stack with tide table data to plan harvest at low tide for maximum access to intertidal species.
fg-04-toxicity-checker— Coastal species pass through the same safety pipeline; Sea Lettuce/Acid Kelp pair flaggedfg-10-preservation-guide— Air-drying methods for Dulse and other seaweeds (4 steps, 1–2 year shelf life)fa-06-marine-life-guide— Marine biodiversity context for the same coastal zone
Tracks 12 wild nut and berry species (8 berries, 4 nuts) by location and season, each with hardcoded ripeness months, edibility ratings, and hazard warnings. Species include Blackberry, Wild Raspberry, Bilberry, Elderberry, Hawthorn, Sloe, Hazelnut, Walnut, Sweet Chestnut, and Oak (acorns). In-season species are highlighted; sort order places currently-ripe species first.
lat, lng (required); radius in km (optional, default 50); month (optional, defaults to current)
Nut and berry species with ripeness months, edibility rating, hazard notes, observation count, photos. In-season-now subset. Safety disclaimer.
iNaturalist v1 (occurrence enrichment, free). Curated 12-species database with ripeness calendars (static). PFAF edibility cross-reference. Trefle (supplemental). 7-day cache. Status: dev.
7 days. Ripeness timing is hardcoded from regional data; iNaturalist enrichment adds observation evidence.
A family plans a berry-picking outing near Hood River, Oregon in August. fg-07 shows 5 species in season: Blackberry (ripeness Aug–Sep, edibility 5), Wild Raspberry (Jul–Aug, edibility 5), Elderberry (Aug–Sep, edibility 4), Hawthorn (Sep–Oct, edibility 3), and Hazelnut (Sep–Oct, edibility 4). Sweet Chestnut triggers a hazard warning: “Do not confuse with Horse Chestnut (Aesculus hippocastanum), which is toxic.” Raw acorns carry a tannin warning requiring leaching. Stack with fg-05-urban-foraging to find mapped berry bushes along accessible trails.
fg-04-toxicity-checker— Sweet Chestnut/Horse Chestnut and Bilberry/Deadly Nightshade pairs flaggedfg-03-seasonal-harvest— Broader seasonal context for nut and berry ripeness within the foraging calendarfg-02-wild-edible-atlas— Wild edible atlas feeds species occurrence data for berries and nutsfg-10-preservation-guide— Jam-making instructions for berries; drying methods for nuts
Surfaces wild plants with documented medicinal properties near a location, drawing from a curated 10-species database with phytochemical profiles, traditional uses, and preparation methods. Sources include Plants For A Future (PFAF) and Dr. Duke’s Phytochemical and Ethnobotanical Database. Species range from Stinging Nettle (iron, anti-inflammatory) through St. John’s Wort (significant drug interactions flagged) to Meadowsweet (natural aspirin precursor—avoid if aspirin-allergic). Traditional uses only, not medical advice.
lat, lng (required); radius in km (optional, default 50); species (optional, single-species lookup)
Medicinal plant list with phytochemicals, traditional uses, preparation methods (tea, tincture, poultice), medicinal rating, hazard warnings, drug interaction flags. Safety disclaimer.
PFAF (edibility + medicinal ratings, static). Dr. Duke’s Phytochemical DB (compounds + ethnobotany, static). iNaturalist v1 (occurrence enrichment, free). 30-day cache. Status: dev.
30 days. Medicinal and phytochemical data is reference-grade; long cache is appropriate.
An herbalist surveys wild medicinal resources near Ashland, Oregon. fg-08 returns 7 species within 50km, led by Stinging Nettle (medicinal rating 5, uses: anti-inflammatory, diuretic, iron supplement) and Elderberry (medicinal rating 4, uses: immune support, cold/flu). St. John’s Wort (Hypericum perforatum) returns with a prominent drug interaction warning: “Interacts with SSRIs, blood thinners, birth control, and immunosuppressants.” Meadowsweet flags an aspirin-allergy contraindication. Stack with fg-04-toxicity-checker for full compound analysis and fg-10-preservation-guide for tincture preparation instructions (alcohol tincture, 6 steps, 3–5 year shelf life).
fg-04-toxicity-checker— Compound-level toxicity and drug interaction data for every medicinal speciesfg-02-wild-edible-atlas— Many medicinal plants are also edible; atlas provides occurrence contextfg-10-preservation-guide— Tincture, tea, and poultice preparation instructions for harvested medicinalsfg-03-seasonal-harvest— Optimal harvest timing for medicinal potency (e.g., nettle before flowering)
Scores current and forecast weather conditions against species-specific foraging requirements using a 100-point system. Evaluates 10 foraging targets (Chanterelle, Morel, Blackberry, Elderberry, Nettle, Dandelion, Ramp, Fiddlehead, Seaweed, Wild Garlic) across temperature, humidity, recent rainfall, wind speed, and seasonality. Each target gets a GO (≥70), MARGINAL (≥40), or WAIT (<40) recommendation. Identifies the best foraging day in the 7-day forecast.
lat, lng (required)
10 foraging targets with weather scores (0–100), go/marginal/wait recommendation. Best day in 7-day forecast. Overall rating (excellent/good/fair/poor). Safety disclaimer.
Open-Meteo forecast API (free, no auth). Includes 3 days historical rain + 7-day forecast. 1-hour cache. Status: dev.
1 hour. Weather changes rapidly; hourly refresh ensures foraging recommendations reflect current conditions.
A forager near Olympia, Washington checks fg-09 on a Thursday morning. Current conditions: 14°C, 88% humidity, 3 rain days in the past week, light wind (8 km/h). Chanterelle scores 85/100 (GO): perfect post-rain temperature and humidity window. Morel scores 35/100 (WAIT): wrong season (April–May). Blackberry scores 72/100 (GO): in-season and warm enough. Best day: Saturday (average score 78 across in-season targets, temperature rising to 17°C). Stack with fg-01-mushroom-finder to identify the specific chanterelle species reported nearby, then fg-10-preservation-guide for drying instructions if the haul exceeds immediate use.
fg-01-mushroom-finder— Weather scoring is especially critical for mushrooms, which fruit 2–5 days after rainfg-03-seasonal-harvest— Season data determines which targets are eligible for weather scoringa-01-current-conditions— Atmos current weather data supplements Open-Meteo forecasta-02-seven-day-forecast— Extended forecast enables multi-day foraging trip planning
A pure static reference rill with no external API calls. Contains 9 curated preservation entries covering 7 species and 6 methods: air drying (Porcini, Dulse), dehydrator drying (Chanterelle), jam-making (Blackberry), syrup (Elderberry, Rose Hip), alcohol tincture (Stinging Nettle), lacto-fermentation (Wild Garlic), sauté-and-freeze (Chanterelle), and cold smoking (Oyster Mushroom). Each entry includes step-by-step instructions, shelf life, and safety notes.
species (optional, common or scientific name); method (optional filter: drying/tincture/fermentation/smoking/jam/syrup)
Preservation entries with method name, step-by-step instructions (4–6 steps), shelf life, safety notes, equipment needed. Methods breakdown by frequency.
Curated static dataset (9 entries, no external dependencies). No TTL (immutable reference data). Status: dev.
None. Static reference data updated only when new preservation methods or species are added to the curation set.
A forager returns from the Tillamook forest with 3 pounds of Golden Chanterelles. fg-10 returns two preservation methods for Cantharellus cibarius: Dehydrator Drying (5 steps, slice 5mm thick, dehydrate at 50°C for 6–8 hours, store in airtight container, 1-year shelf life) and Sauté & Freeze (5 steps, sauté in butter until moisture released, cool, vacuum-seal, freeze, 6–12 month shelf life). The forager chooses to dry half and freeze half. Stack with fg-01-mushroom-finder and fg-09-weather-window upstream for the complete foray-to-pantry pipeline.
fg-01-mushroom-finder— Identifies what you found; Preservation Guide tells you how to keep itfg-02-wild-edible-atlas— Wild edible identification feeds into preservation method selectionfg-03-seasonal-harvest— Seasonal glut timing drives preservation planning
cultura (cultivation, tended growth)
If Forage asks “what grows wild here?” then Cultura asks “what should I plant here?” This is the cultivated species layer—vegetables, herbs, fruits, and grains in managed garden systems. It spans growing calendars anchored to USDA hardiness zones, crop databases from OpenFarm and Trefle, companion planting matrices, seed library discovery, soil requirement matching against USDA Soil Data Access, pest and disease identification, permaculture guild design, and harvest prediction. Several rills are heavily interdependent: soil suitability (cu-05) informs crop selection (cu-02), which informs companion planting (cu-03), which feeds permaculture design (cu-07). This is a hybrid API-plus-static family where curated datasets (permaculture guilds, seed libraries) are as important as runtime API calls.
Planting and harvest date windows by hardiness zone, crop growing requirements (sun, water, spacing, days to maturity), beneficial and harmful plant pairings, seed library and seed bank locations, crop-specific soil pH and amendment recommendations, garden pest and disease identification with organic treatments, permaculture guild compositions and food forest layer assignments, and harvest date prediction with succession planting windows.
OpenFarm API v1 (crop data, free). Trefle API (botanical specs, token required). USDA Soil Data Access / SSURGO (soil profiles, free). USDA PHZM (hardiness zones, free). Perenual API (pest/disease data, key required). iNaturalist v1 (regional pest observations, free). Curated static datasets for companion planting matrix, permaculture guilds (5 guilds from Hemenway, Jacke & Toensmeier, Mollison), seed libraries (10 locations), and yield estimates (10 crops).
A planting calendar alone is a guess. Stack cu-01-growing-calendar with f-01-hardiness-zone and the calendar is anchored to your actual USDA zone with computed frost dates. Layer cu-05-soil-requirements with t-02-soil-profile and you get a match score between what a crop needs and what your soil provides, with specific amendment recommendations. Combine cu-03-companion-planting with cu-07-permaculture-design and individual plant pairings scale into full guild compositions. Stack cu-08-harvest-predictor with cu-01 and a single planting date generates a complete timeline from indoor start through succession harvest windows.
Stack cu-01-growing-calendar + cu-02-crop-database + cu-03-companion-planting + cu-05-soil-requirements for a new garden bed in Portland, Oregon (zone 8b). Growing Calendar computes frost dates: last spring March 10, first fall November 15. Crop Database returns tomato specs: full sun, moderate water, 60–85 days to maturity, 60cm spacing. Companion Planting shows tomato + basil (beneficial: pest deterrent, flavor enhancement) and tomato + fennel (antagonist: growth inhibition). Soil Requirements queries USDA SDA for the parcel’s soil profile (Multnomah silt loam, pH 5.8) and computes a 72/100 match score for tomatoes (pH slightly low), recommending lime amendment. The composition transforms a vague “I want to grow tomatoes” into a zone-specific, soil-matched, companion-optimized planting plan.
| ID | Name | Description |
|---|---|---|
| cu-01 | Growing Calendar | Planting and harvest date windows by USDA hardiness zone with frost date calculations |
| cu-02 | Crop Database | Vegetable, herb, and fruit growing requirements from OpenFarm and Trefle |
| cu-03 | Companion Planting Guide | Beneficial and harmful plant pairings with mechanisms from curated matrix and OpenFarm |
| cu-04 | Seed Library Finder | Nearby seed libraries, seed banks, and community gardens with contact info |
| cu-05 | Soil Requirements | Crop-specific soil matching with USDA SSURGO profiles, pH scoring, and amendment recommendations |
| cu-06 | Pest & Disease Identifier | Garden pests and diseases with organic treatment options and regional threat data |
| cu-07 | Permaculture Design Data | Guild compositions, food forest layers, and polyculture groupings from published references |
| cu-08 | Harvest Predictor | Days-to-maturity estimates, yield predictions, and succession planting windows |
Generates planting and harvest date windows by USDA hardiness zone. Combines a curated frost date lookup table (zones 3a–10b) with days-to-maturity data from OpenFarm to produce indoor start, transplant, direct sow, and harvest windows for each crop. Default crops: tomato, basil, carrot, lettuce, pepper, zucchini. Succession intervals computed automatically (14 days for fast crops, 21 for moderate).
lat, lng (required); crop (optional, filters to single crop)
Per-crop calendar entries with indoor start week, transplant week, direct sow week, harvest start/end weeks, succession interval. Zone ID and frost dates (last spring, first fall).
OpenFarm API v1 (crop data, free, 500ms courtesy delay). USDA PHZM (zone lookup). Curated frost date table (zones 3a–10b). 24-hour cache. Status: dev.
24 hours. Frost dates are annual; crop maturity data changes infrequently. Daily refresh catches OpenFarm updates.
A first-time gardener in Portland, Oregon (zone 8b) asks cu-01 when to plant tomatoes. Frost dates: last spring March 10, first fall November 15. OpenFarm returns 75 days to maturity. Calendar computes: indoor start week 8 (late Feb), transplant week 14 (early Apr, 4 weeks after last frost), direct sow week 16, harvest start week 24 (mid-Jun), harvest end week 30 (late Jul). No succession interval (DTM > 90). Stack with f-01-hardiness-zone for zone verification and cu-08-harvest-predictor for yield estimates.
f-01-hardiness-zone— Zone lookup that anchors the frost date calculationcu-08-harvest-predictor— Growing Calendar feeds planting dates into yield and harvest predictioncu-02-crop-database— Crop specs (spacing, sun, water) complement the calendar’s timing datacu-05-soil-requirements— Soil matching ensures the calendar’s crops will actually thrive in local soil
Parallel-fetches crop growing requirements from OpenFarm and Trefle, merging both into a unified crop guide. OpenFarm provides growing data (sowing method, spacing, days to maturity); Trefle supplements with botanical specifications (light tolerance, soil humidity, atmospheric humidity). Light requirements mapped to human-readable scale: 0–3 Full Shade, 4–5 Partial Shade, 6–7 Partial Sun, 8–10 Full Sun. Water needs: 0–3 Low, 4–6 Moderate, 7–10 High.
query (required, crop name); lat, lng (optional, for future zone-specific filtering)
Crop guide with common/scientific names, sun requirements, water needs, spacing, height range, days to maturity range, sowing method, growing tips. Source attribution.
OpenFarm API v1 (growing data, free). Trefle API (botanical specs, token required, 1100ms rate limit). Merged result. 30-day cache. Status: dev.
30 days. Crop reference data is stable; monthly refresh captures community edits on OpenFarm.
A gardener queries cu-02 for “basil.” OpenFarm returns: sowing method indoor/outdoor, spacing 30cm, days to maturity 60–70. Trefle supplements: light tolerance 8 (Full Sun), atmospheric humidity 5 (Moderate), height 30–60cm. The merged guide gives a complete growing profile. Stack with cu-03-companion-planting—basil is a beneficial companion for tomato (pest deterrent, flavor enhancement)—and cu-01-growing-calendar for zone-specific planting dates.
cu-01-growing-calendar— Days-to-maturity data drives planting calendar computationscu-03-companion-planting— Crop identity feeds companion lookup for beneficial/harmful pairingscu-08-harvest-predictor— DTM and yield data feed harvest predictioncu-05-soil-requirements— Crop water and light needs complement soil matching
Returns beneficial and harmful plant pairings for a given crop by merging a curated static matrix (6 crops: tomato, basil, carrot, lettuce, pepper, zucchini) with OpenFarm companion data. Curated entries include mechanisms (“nitrogen fixation,” “pest deterrent,” “allelopathic inhibition”) and get “verified” confidence; OpenFarm additions get “anecdotal” confidence. Deduplicated by companion name.
crop (required, crop name)
Companion pairings (beneficial + antagonist) with companion name, relationship type, mechanism description, confidence level (verified/anecdotal), source. Total pairing count.
Curated companion planting matrix (static, 6 crops). OpenFarm API v1 (companion field, free). Merged and deduplicated. 90-day cache. Status: dev.
90 days. Companion planting data is stable reference knowledge; quarterly refresh captures OpenFarm community edits.
A gardener planning a raised bed queries cu-03 for “tomato.” Beneficial companions: Basil (pest deterrent, flavor enhancement, verified), Carrot (soil aeration, verified), Marigold (nematode suppression, verified). Antagonists: Fennel (allelopathic growth inhibition, verified), Brassicas (nutrient competition, verified). The gardener places basil next to tomatoes and moves fennel to a separate bed. Stack with cu-07-permaculture-design to scale individual pairings into full guild compositions with layer assignments.
cu-02-crop-database— Crop identity and spacing data feed companion planting recommendationscu-07-permaculture-design— Pairings scale into full permaculture guild compositionscu-06-pest-disease-identifier— Companion mechanisms (pest deterrent) complement pest identification data
Locates nearby seed libraries, seed banks, and community garden seed collections from a curated database of 10 locations across the US. Entries include public libraries with seed lending programs (Richmond, SF, Berkeley, Portland, Chicago, Austin), dedicated seed banks (Hudson Valley, National Center for Genetic Resources Preservation, Seed Savers Exchange), and community garden collections (Denver Urban Gardens). Results sorted by distance using Haversine formula.
lat, lng (required); radius in km (optional, default 50)
Seed library list with name, type (library/seed-bank/community-garden), address, distance in km, website, last-verified date. Count and search radius.
Curated static dataset (10 entries, no external API). 365-day cache (effectively immutable). Status: dev.
Annual. Seed libraries are stable institutions; annual verification of addresses and URLs is sufficient.
A gardener in inner SE Portland queries cu-04. One result within 50km: Portland Seed Library (3.2km away, type: library, website: portlandseedlibrary.net, last verified Aug 2025). The nearest seed bank is Seed Savers Exchange in Decorah, Iowa (2,900km). The gardener visits the Portland Seed Library to borrow heirloom tomato seeds, then uses cu-01-growing-calendar for zone 8b planting dates and cu-02-crop-database for growing requirements.
cu-01-growing-calendar— Borrowed seeds need zone-specific planting datescu-02-crop-database— Growing requirements for borrowed cultivarscu-03-companion-planting— Companion recommendations for seed library selections
Matches crop-specific soil needs against actual local soil conditions by querying two sources in parallel: USDA Soil Data Access (SSURGO database via SQL, returns horizon-level pH, organic matter, sand/clay/silt percentages, drainage class, and texture) and Trefle (crop-specific pH range, nutrient needs, salinity tolerance). Computes a 0–100 match score with deductions for pH mismatch (up to −40), low organic matter (−15), and poor drainage (−15). Generates specific amendment recommendations.
crop (required); lat, lng (required)
Local soil profile (pH, OM%, sand/clay/silt, drainage, texture class), crop soil needs (pH range, nutrient level, humidity, salinity tolerance), match score (0–100), amendment recommendations.
USDA Soil Data Access / SSURGO (SQL-based, free). Trefle API (botanical specs, token required). 365-day cache. Status: dev.
Annual. SSURGO soil data is survey-based and updates on multi-year cycles. Local soil doesn’t change rapidly.
A gardener in Beaverton, Oregon checks whether blueberries will thrive in their backyard. cu-05 queries SSURGO: Multnomah silt loam, pH 5.8, organic matter 3.2%, well-drained. Trefle returns blueberry needs: pH 4.5–5.5, high acidity required. Match score: 62/100 (pH 0.3 too high, −20 deduction). Amendments recommended: “Add sulfur to lower pH by 0.3–0.5 units; incorporate peat moss for acidity and organic matter.” Stack with cu-02-crop-database for full growing requirements and t-02-soil-profile for deeper soil characterization.
cu-02-crop-database— Crop identity feeds soil requirements lookupt-02-soil-profile— Terra Firma soil data provides complementary pedological contextcu-01-growing-calendar— Soil-matched crops feed into zone-specific planting calendarscu-07-permaculture-design— Soil suitability informs guild and polyculture species selection
Identifies common garden pests and diseases for a given crop by merging three sources: a curated static database (tomato: 3 entries, pepper: 2, lettuce: 2), Perenual API pest/disease listings, and iNaturalist regional insect observations within 25km. Each pest includes organic treatment options drawn from a curated treatments map covering aphids, whiteflies, spider mites, hornworms, powdery mildew, blight, root rot, slugs, and flea beetles. Deduplicated by pest name.
crop (required); lat, lng (optional, enables regional threat detection)
Pest and disease list with name, type (pest/disease/fungal), severity, organic treatments, prevention tips. Regional insect threats with observation counts. Total threat count.
Perenual API (pest/disease data, key required). iNaturalist v1 (regional Insecta observations, free). Curated pest database + organic treatment map (static). 30-day cache. Status: dev.
30 days. Pest populations shift seasonally; monthly refresh captures regional outbreak patterns.
A gardener in the Willamette Valley notices holes in their tomato leaves and queries cu-06 for “tomato.” Results: Tomato Hornworm (severity: high, treatment: “hand-pick, Bt spray, encourage parasitic wasps”), Aphids (severity: moderate, treatment: “neem oil spray, encourage ladybugs, strong water spray”), Early Blight (severity: high, treatment: “remove affected foliage, copper spray, improve air circulation”). Regional threats from iNaturalist: Japanese Beetle (47 observations within 25km)—not yet on the tomato pest list but present in the area. Stack with cu-03-companion-planting for pest-deterrent companions (basil deters hornworms).
cu-02-crop-database— Crop identity feeds pest lookupcu-03-companion-planting— Companion plants with pest-deterrent mechanisms complement organic treatmentsfa-09-insect-atlas— Broader insect diversity context for the same region
A pure static reference rill with no external API calls. Contains 5 curated permaculture guilds from published sources (Hemenway’s Gaia’s Garden, Jacke & Toensmeier’s Edible Forest Gardens, Mollison, Traditional Indigenous agriculture) and 8 food forest layer definitions (canopy through mycelial). Guilds: Apple Tree Guild, Three Sisters, Berry Patch, Mediterranean Herb Spiral, Nitrogen Fixer Support. Each guild includes zone compatibility, center plant, member species with layer assignments and roles (nitrogen fixer, dynamic accumulator, pest deterrent, pollinator attractor).
guild (optional, guild name or center plant); zone (optional, hardiness zone filter)
Guild compositions with center plant, member species (with layer and role), zone compatibility, source reference. 8 food forest layer definitions with height ranges and example plants.
Curated static dataset (5 guilds, 8 layers). No external dependencies. 365-day cache (effectively immutable). Status: dev.
Annual. Permaculture guild data is reference-grade from published books; updated only when new guild designs are added.
A homesteader in the Willamette Valley (zone 8b) queries cu-07 for guilds compatible with their zone. Three guilds match: Apple Tree Guild (zones 4a–8b, members: Apple, Comfrey, White Clover, Daffodil, Chives, Nasturtium, Yarrow), Berry Patch (zones 4a–7b, marginal), and Nitrogen Fixer Support (zones 4a–8a). The Apple Tree Guild assigns each member to a food forest layer: Apple (canopy), Comfrey (herbaceous, dynamic accumulator), White Clover (ground cover, nitrogen fixer), Daffodil (herbaceous, pest deterrent). Stack with cu-03-companion-planting to verify individual pairings within the guild and cu-05-soil-requirements to check soil suitability for the center plant.
cu-03-companion-planting— Validates individual pairings within guild compositionscu-05-soil-requirements— Soil suitability check for guild center plantsf-01-hardiness-zone— Zone lookup determines which guilds are compatible with a given locationcu-01-growing-calendar— Zone-specific timing for planting guild members
Given a crop name, planting date, and location, predicts harvest windows and generates succession planting schedules. Days-to-maturity fetched from OpenFarm (with ±15% range). Yield estimates from a curated 10-crop database (tomato 4.5–9kg per plant, lettuce 0.2–0.5kg, corn 0.5–1 ear). Succession intervals for 9 quick-cycle crops (lettuce every 14 days, radish every 10). Season end approximated by latitude: >40°N ends Sep, >35°N ends Oct, else Nov.
crop (required); plantingDate (required, ISO 8601); lat, lng (required)
Harvest estimate with start/end dates, DTM range, yield per plant. Succession windows (up to 8) with sow dates and expected harvest dates. Confidence level (high if yield data known, medium otherwise).
OpenFarm API v1 (DTM data, free). Curated yield and succession interval datasets (static). 30-day cache. Status: dev.
30 days. DTM data is stable; monthly refresh captures OpenFarm community corrections.
A gardener in Portland (45.5°N, season end: September) plants lettuce on April 1. cu-08 returns: DTM 45–55 days, harvest window May 16–Jun 9, yield 0.2–0.5kg per plant. Succession schedule: 7 windows at 14-day intervals (Apr 1, Apr 15, Apr 29, May 13, May 27, Jun 10, Jun 24, Jul 8), last harvest Aug 31. The gardener now has a complete sowing-to-harvest timeline for continuous lettuce through summer. Stack with cu-01-growing-calendar for frost-date validation and cu-05-soil-requirements to confirm the bed’s soil suits lettuce.
cu-01-growing-calendar— Planting dates from the growing calendar feed harvest predictioncu-02-crop-database— DTM and crop specs provide the prediction inputscu-05-soil-requirements— Soil match quality may affect actual yield vs. predicted yield
Polis — The Built & Governed World
Once you know what the ground is made of and what lives on it, the next question is: what have humans done here? Polis is the regulatory and infrastructural layer—zoning codes, property values, energy grids, road networks, parcel boundaries, and movement patterns. This is the domain where nature meets policy, where soil type becomes buildability and sunlight becomes solar potential on a permitted rooftop.
These rills pull from Regrid, Census Bureau, FEMA, EIA, NREL, OpenStreetMap, Mapbox, and dozens of city-level open data portals. Individually, a zoning code is a bureaucratic fact. Stacked with terrain, ecology, and market data, it becomes a decision surface—a map of what’s possible.
civitas (citizenship, community, the state)
Who governs this place, and what are the rules? Civitas answers the questions that anyone buying, building on, or investing in a property needs answered first: what’s the zoning, what permits have been issued, what are the setback requirements, who represents this district, and what does the municipal code say about that fence you want to build. It reaches from parcel-level zoning data up through school districts, voting districts, and elected officials—the full civic stack for any US address.
Zoning designations and permitted uses, building permits, municipal ordinances, property setbacks and development regulations, elected officials, school district boundaries, voting districts, property tax rates, land use history, and code violations.
Regrid Parcel API (zoning/parcels, paid key), Socrata/SODA (permits/violations, free), Municode (ordinances, free/unofficial), OpenStates API (legislators, free key), Census Bureau Geocoder (districts, free), CourtListener (court cases, free key), Nominatim/OSM (geocoding, free).
A zoning code alone is a bureaucratic designation. Stack cv-01-zoning-lookup with t-02-soil-profile and gr-03-solar-potential, and it becomes an answer to “can I build a solar-powered greenhouse here?” Combine cv-04-property-regulations with vl-01-property-valuation and vl-02-flood-risk, and you surface whether a property’s price reflects its actual development constraints. Layer cv-06-school-district with cv-08-tax-rate and vl-08-cost-of-living, and you’ve built a family relocation scorecard that no single dataset can provide.
Stack cv-01-zoning-lookup + cv-04-property-regulations + cv-02-permit-tracker + vl-01-property-valuation + t-02-soil-profile for a parcel in Seattle’s Ballard neighborhood. The zoning says LR1 (low-rise residential), the setbacks require 5-foot side yards, permits show two ADU conversions on the same block in the last year, the valuation is $485K, and the soil is well-drained sandy loam. This surfaces a signal that the neighborhood is actively densifying under current zoning—the permit activity combined with favorable soil and modest setbacks suggests ADU potential that the property price may not yet fully reflect.
| ID | Name | Description |
|---|---|---|
| cv-01 | Zoning Lookup | Current zoning designation, type, subtype, and permitted uses via Regrid |
| cv-02 | Permit Tracker | Active and recent building permits from city open data portals via Socrata |
| cv-03 | Municipal Code Search | Full-text search of local ordinances via Municode covering 3,500+ municipalities |
| cv-04 | Property Regulations | Setbacks, FAR, height limits, coverage, and density restrictions from Regrid zoning |
| cv-05 | Elected Officials | State and federal legislators with party, district, and contact info via OpenStates |
| cv-06 | School District Profile | School district name, type, and GEOID via Census Bureau Geocoder |
| cv-07 | Voting District Map | Congressional, state senate, and assembly districts via Census Geocoder |
| cv-08 | Tax Rate Lookup | Property tax rates at state level from Census ACS data with Nominatim geocoding |
| cv-09 | Land Use History | Historical zoning and NLCD satellite land cover changes over time |
| cv-10 | Legal Compliance Check | Code violations from Socrata and court cases from CourtListener |
Identifies the current zoning designation for any US parcel—code, type, subtype, and permitted uses—using Regrid’s standardized 26-field zoning schema. The first question anyone should ask about a piece of land, and usually the last one they think of.
lat, lng (required)
Zoning code, description, type, subtype, objective, permitted uses grouped by category, parcel address, APN, and owner
Regrid Parcel API v2. Paid (subscription after 7-day trial). Key required.
Monthly. Zoning changes are rare—30-day cache is appropriate.
A couple looking at a half-acre lot near Beaverton, Oregon queries cv-01 and discovers it’s zoned R-5 (residential, 5,000 sq ft minimum lot size). The permitted uses include single-family dwelling and accessory dwelling unit. Stack with cv-04-property-regulations to learn the specific setback and height limits, then with cv-02-permit-tracker to see if neighbors have already built ADUs under this zoning—a signal that the regulatory path is well-worn.
cv-04-property-regulations— Adds dimensional limits (setbacks, FAR, height) to the zoning designationvl-01-property-valuation— Reveals whether the property is priced for its zoned use or for speculative upzoningt-02-soil-profile— Soil drainage class modulates what “buildable” actually means for the zoned usecv-02-permit-tracker— Recent permits show how the zoning is being exercised in practicefm-06-fundus— Adds parcel geometry and ownership to the zoning context
Tracks active and recent building permits near any US location by querying city-specific open data portals via Socrata. Covers Chicago, Seattle, San Francisco, Los Angeles, and New York—with the architecture ready for more cities. Permits are the exhaust trail of a neighborhood changing.
lat, lng (required); radiusMeters (optional, default 500)
Array of permits with type, description, address, cost, issue date, and status; resolved city name; total count
Socrata/SODA API (spatial queries). Free, 1,000 req/hr with optional app token. Nominatim for reverse geocoding.
Daily. Permits update frequently—24-hour cache is appropriate.
An investor evaluating a fourplex in Chicago’s Logan Square runs cv-02 within a 500-meter radius and finds 14 permits issued in the last 6 months—8 of them for “interior renovation” and 3 for “new construction.” Combined with vl-07-market-trends, this surfaces a gentrification signal: the neighborhood is investing in upgrades and new builds, which may indicate rising property values but also increasing construction noise and displacement pressure.
vl-07-market-trends— Permit velocity correlates with market appreciation—or overheatingcv-01-zoning-lookup— Permits contextualized by what the zoning allows reveals neighborhood directioncv-10-legal-compliance— Cross-reference permits with violations to detect unpermitted workc-06-points-of-interest— Construction permits near parks or trails may signal development pressure on green space
Full-text search of local ordinances and municipal codes across 3,500+ US municipalities via the Municode database. Finally, an answer to “is it legal to keep chickens in my backyard?” that doesn’t require a law degree or an afternoon of Googling.
lat, lng (required); query (required, e.g. “fence height”, “ADU”, “setback”)
Search results with section headings, text snippets, node IDs, and direct Municode Library links; resolved municipality name; total results count
Municode API (unofficial/undocumented). Free, no auth. Nominatim for reverse geocoding. Rate limits unknown.
Monthly. Ordinances update periodically—30-day cache is appropriate.
A homeowner in Portland queries cv-03 with “accessory dwelling unit” and gets back 12 relevant code sections including Title 33.205 (Accessory Dwelling Units), with snippets showing the specific dimensional standards and owner-occupancy requirements. Stack with cv-04-property-regulations to see if the parcel’s actual setbacks and lot area meet the ordinance requirements—a composition that moves from “what does the law say” to “does my lot qualify.”
cv-04-property-regulations— Code text + dimensional data = a compliance check rather than just a reading assignmentcv-01-zoning-lookup— Municipal code references zoning designations that this rill resolvescv-02-permit-tracker— Code sections + recent permits show whether the regulation is actively enforced
The dimensional rulebook for a parcel: front/rear/side setbacks, maximum height, floor area ratio, density limits, and coverage percentages from Regrid’s standardized zoning schema. Eleven specific regulation fields extracted and parsed, including Regrid’s sentinel values for “consult local code” and “not applicable.”
lat, lng (required)
Full regulation set with 11 parsed values (setbacks, FAR, height, coverage, density), permitted uses, zoning code, parcel address
Regrid Parcel API v2 (same endpoint as cv-01). Paid key required.
Monthly. Zoning regulations are stable—30-day cache is appropriate.
A designer planning a backyard studio in San Francisco’s Outer Sunset queries cv-04 and finds: 5-foot side setback, 25-foot rear setback, 40-foot max height, and 0.55 FAR. The lot is 2,500 sq ft, meaning the total allowable floor area is 1,375 sq ft. The existing house is 1,100 sq ft, leaving 275 sq ft for the studio. Stack with t-01-helio-study to determine if the studio placement within the setback envelope gets adequate natural light.
cv-01-zoning-lookup— Regulations contextualized by zoning designation explain the “why” behind the numberst-01-helio-study— Setback envelope + solar position = buildable area with natural light analysisfm-05-vestigium— Building footprints show how neighbors have used their setback envelopesvl-01-property-valuation— Development constraints directly affect property value
Identifies state and federal legislators for any US location via the OpenStates API. Returns name, party, district, contact info, and photo. Does not include local officials (city council, mayor)—that’s a gap no free API fills well.
lat, lng (required)
Array of officials with name, party (D/R/I), title, chamber, district, email, phone, website, office address, and photo URL; state jurisdiction name
OpenStates API v3. Free tier (~500 queries/day). API key required.
Weekly. Legislators change infrequently—7-day cache is appropriate.
A community organizer in Tacoma, Washington queries cv-05 and finds their state senator (D-27th District), two state representatives, and their US congressional representative. Stack with cv-07-voting-district to map these officials to their exact district boundaries, and with cv-03-municipal-code to find which ordinances they’d need to advocate changing—a composition that moves from “who represents me” to “who do I talk to about this specific rule.”
cv-07-voting-district— Officials mapped to their district boundaries for spatial contextcv-03-municipal-code— Know who to contact about which ordinancecm-03-civic-engagement— Representative info combined with community engagement data
Identifies the school district for any US location using the Census Bureau Geocoder. Returns district name, GEOID, and type (unified, elementary, or secondary). Free government data, no API key, and it handles the edge case of overlapping elementary and secondary districts.
lat, lng (required)
Array of school districts (can be multiple if elementary + secondary overlap), county name, district GEOID and type
Census Bureau Geocoder (geographies/coordinates). Completely free, no API key required.
Annually. School district boundaries are very stable—365-day cache is appropriate.
A family relocating to the Portland metro area queries cv-06 for several neighborhoods and finds that a house in Lake Oswego falls in the Lake Oswego School District (unified), while one in unincorporated Clackamas County falls in the North Clackamas School District. Stack with cv-08-tax-rate to see how district boundaries affect property tax obligations, and with vl-08-cost-of-living to understand the full financial picture of each location.
cv-08-tax-rate— School district funding often explains tax rate differences between adjacent parcelsvl-01-property-valuation— School district quality is a well-documented driver of property valuesvl-08-cost-of-living— District + tax + cost data composes into a relocation analysisc-05-boundaries— District boundaries on a map show where jurisdictions change
Identifies congressional districts, state senate and assembly districts, and county for any US location. Currently returns district IDs via Census Geocoder; polygon geometry via TIGERweb is architecturally planned but not yet wired.
lat, lng (required)
Congressional district, state senate district, state assembly district, and county—each with name, GEOID, and district number
Census Bureau Geocoder (119th Congressional, 2024 State Legislative layers). Free, no auth. Planned: TIGERweb for polygon geometry.
Annually (post-redistricting). 365-day cache is appropriate.
A nonprofit analyzing civic engagement in Seattle queries cv-07 and finds the location falls in Washington’s 7th Congressional District, 43rd Legislative District (state senate and house). Stack with cv-05-elected-officials to immediately identify the specific representatives, and with cm-01-neighborhood-profile to understand the demographic context—a composition that connects political geography to the people it represents.
cv-05-elected-officials— District IDs resolve directly to the people who hold those seatscv-06-school-district— Overlapping political and school district boundaries reveal governance complexityc-05-boundaries— Cartographic rendering of district polygons on a shared map
Looks up property tax rates for any US location. Currently uses a static state-level lookup table derived from Census ACS data with Nominatim reverse geocoding. Provides 25th, 50th, and 75th percentile rates—useful for ballpark comparisons, though county-level variation is not yet captured.
lat, lng (required)
Effective property tax rates at 25th/50th/75th percentiles, jurisdiction label, ZIP, state, county, city
Static state-level Census ACS data (built-in lookup table, all 50 states + DC). Nominatim for geocoding. Free, no auth.
Quarterly. Tax rates change annually—90-day cache is appropriate.
A remote worker comparing property tax burdens between Portland (Oregon) and Vancouver (Washington) queries cv-08 for both. Oregon shows a median effective rate of 0.93%; Washington shows 0.84%. But stack with vl-08-cost-of-living and the picture shifts—Oregon has no sales tax, so the property tax comparison alone is misleading. This surfaces a signal that tax comparisons require the full cost-of-living context.
vl-08-cost-of-living— Tax rate in the context of overall living costs reveals the real burdenvl-01-property-valuation— Tax rate × property value = annual tax bill estimatecv-06-school-district— Tax rate differences often correlate with school district quality and funding
Tracks historical zoning and land use changes for a parcel. Combines current zoning from Regrid with USGS NLCD satellite land cover data across 9 observation years (2001–2021). Note: the NLCD component currently uses simulated data based on current zoning; real NLCD WMS integration is planned.
lat, lng (required)
Current zoning summary, NLCD land cover timeline (2001–2021 at 30m resolution), parcel ownership history events
Regrid Parcel API (zoning, paid key). Planned: USGS NLCD WMS/WCS (satellite, free). Currently NLCD is simulated from zoning type.
Monthly. Historical data is stable—30-day cache is appropriate.
An environmental consultant evaluating a parcel near Bend, Oregon queries cv-09 and sees that the NLCD timeline shows the site classified as “Evergreen Forest” from 2001 through 2016, then “Developed, Low Intensity” from 2019 onward. Combined with t-05-land-use for the current NLCD class and vl-02-flood-risk for hydrological context, this surfaces a question about impervious surface increase and stormwater management that a purely present-tense analysis would miss.
t-05-land-use— Current NLCD class compared to historical timeline reveals land use trajectoryvl-02-flood-risk— Deforestation history modulates flood risk assessmentcv-01-zoning-lookup— Current zoning vs. historical land use reveals development pressure
Checks for code violations, enforcement actions, and related court cases near a location. Combines city-specific violation records from Socrata (San Diego, LA, Buffalo, Montgomery County) with court case search from CourtListener. Both sources soft-fail gracefully when data isn’t available.
lat, lng (required); query (optional, default “zoning violation”)
Code violations with type, address, status, and dates; court cases with name, court, filing date, snippet, and citation; resolved city name
Socrata/SODA (violations, free), CourtListener API v4 (court cases, free key, 5,000 req/day), Nominatim (geocoding, free).
Weekly. Violation records update regularly—7-day cache is appropriate.
A buyer considering a commercial property in San Diego queries cv-10 with “building code violation” and finds 3 open violations within 500 meters, including one for “unpermitted construction” at the subject address. The CourtListener search returns a 2023 case in the same jurisdiction involving a similar violation type. Stack with cv-02-permit-tracker to check whether the violation has since been resolved by a permit—a composition that moves from “there was a problem” to “was it fixed.”
cv-02-permit-tracker— Cross-reference violations with permits to see if issues were resolvedvl-01-property-valuation— Unresolved violations may suppress property valuecv-01-zoning-lookup— Violation type + zoning context reveals whether it’s a use conflict or structural issue
valorem (value, worth, price)
What is this place worth, and what are the risks that could change that number? Valorem covers the full financial profile of a property: automated market valuation, comparable sales analysis, tax assessments, flood risk and insurance exposure, rental market data, market trend tracking, and cost-of-living indexing. It reaches from the hyper-local (what did the house next door sell for?) to the regional (how does this area’s housing cost compare to the national average?). The family runs primarily on the RentCast API for property-level data, supplemented by FEMA, USGS, HUD, BLS, and Census datasets for risk and economic context.
Property valuation and AVM estimates, comparable sales, tax assessments, flood zone designations and NFIP claims history, composite insurance risk scores, rental market analysis and Fair Market Rents, market price trends and inventory, cost-of-living indices by category.
RentCast API (valuation, comps, tax, rent, trends; paid key, 50 req/mo free tier), FEMA NFHL MapServer (flood zones, free), OpenFEMA (NFIP claims, free), USGS Earthquake Hazards API (seismic, free), HUD FMR API (Fair Market Rents, free key), BLS CPI API v1 (consumer prices, free), Census ACS 5-year (housing costs and income, free), Nominatim (geocoding, free).
A property valuation in isolation is a number. Stack vl-01-property-valuation with cv-01-zoning-lookup and cv-04-property-regulations, and it becomes an answer to “is this property priced below what its zoning allows?” Combine vl-02-flood-risk with vl-05-insurance-risk and t-02-soil-profile, and you surface whether a bargain price is actually reflecting hidden environmental exposure. Layer vl-06-rental-market with vl-07-market-trends and vl-08-cost-of-living, and you’ve built an investment viability assessment that no single MLS listing can provide.
Stack vl-01-property-valuation + vl-04-comparable-sales + vl-02-flood-risk + vl-06-rental-market + vl-03-tax-assessment for a duplex in Portland’s St. Johns neighborhood. The AVM returns $425K with high confidence from 8 comps, the flood zone is X (minimal hazard), rental comps show $1,850/month for a 2BR, and the effective tax rate is 1.12%. The rent-to-price ratio of 0.44% monthly suggests modest cash flow—but the comps reveal prices climbing 6.2% YoY, and the flood zone means no mandatory insurance surcharge. The composition surfaces a signal that cash flow is thin but appreciation trajectory and low risk exposure make the basis solid.
| ID | Name | Description |
|---|---|---|
| vl-01 | Property Valuation | Automated market value estimate (AVM) with price range and confidence via RentCast |
| vl-02 | Flood Risk Assessment | FEMA flood zone designation, SFHA status, BFE, and county-level NFIP claims history |
| vl-03 | Tax Assessment | Assessed value, annual tax amount, effective tax rate, and property attributes via RentCast |
| vl-04 | Comparable Sales | Recent comparable property sales with price per square foot, distance, and correlation score |
| vl-05 | Insurance Risk Score | Composite 1–10 risk score from flood, seismic, wildfire, and wind hazard layers |
| vl-06 | Rental Market Analysis | Property-level rent estimates, HUD Fair Market Rents, and rental comps with rent-to-price ratio |
| vl-07 | Market Trend Tracker | ZIP-level median sale price, average rent, days on market, and 12-month historical trends |
| vl-08 | Cost of Living Index | Local cost-of-living index relative to national average from BLS CPI and Census ACS data |
Returns an automated market value estimate for any US property—estimated value, price range, confidence score, and the comparable sales that informed the model. The number everyone wants first, and the one that means nothing without context.
lat, lng (required)
Estimated value, price range (low/high), confidence score (0–1), price per sqft, formatted display values, and array of comparable properties with sale prices and correlation scores
RentCast API /v1/value-estimate. Keyed (RENTCAST_API_KEY). 50 req/month free tier. Reverse-geocodes via Nominatim first.
Weekly. 7-day cache reflects typical AVM update cycles.
A buyer evaluating a listing in Bend, Oregon queries vl-01 and gets an AVM of $512K with a range of $485K–$540K at 0.82 confidence. The listing asks $529K—within range but above estimate. Stack with vl-04-comparable-sales to see that the 8 comps averaged $225/sqft while the listing is $240/sqft, then with vl-07-market-trends to see if prices are accelerating or plateauing in this ZIP. The AVM alone doesn’t tell you whether $529K is fair; the composition does.
vl-04-comparable-sales— Surfaces the individual comps that drive the AVM numbervl-03-tax-assessment— Reveals gap between market value and assessed value (shares RentCast call budget)vl-06-rental-market— Computes rent-to-price ratio for investment analysiscv-01-zoning-lookup— Connects price to what the land is permitted to become
Identifies the FEMA flood zone for any US location and enriches it with historical NFIP claims data. Returns zone designation, Special Flood Hazard Area status, Base Flood Elevation, risk tier, and county-level claims history going back a decade. The insurance bill you didn’t budget for.
lat, lng (required)
Flood zone code (e.g., AE, X), SFHA boolean, Base Flood Elevation and datum, risk tier (High/Moderate/Low/Undetermined), county-level NFIP claims with annual breakdown
FEMA NFHL MapServer Layer 28 (ArcGIS REST, free, no auth) + OpenFEMA /v2/FimaNfipClaims (free, no auth) + Nominatim. Flood/county queries run in parallel.
Annually. Flood zone maps rarely change—365-day cache is appropriate.
A property near the Willamette River in Portland returns Zone AE (1% annual chance flood hazard) with SFHA status true and a BFE of 28 feet NAVD88. County claims history shows 847 total NFIP claims with $12.3M paid out, averaging $14,500 per claim. Stack with vl-05-insurance-risk for the composite risk picture, then with vl-01-property-valuation to evaluate whether the asking price reflects the mandatory flood insurance premium that SFHA status requires.
vl-05-insurance-risk— Flood score feeds into the composite insurance risk calculationvl-01-property-valuation— Connects flood exposure to property pricet-02-soil-profile— Soil drainage characteristics affect actual flood behavior beyond FEMA zones
Retrieves the tax assessment for a US property—assessed value split between land and improvements, annual tax amount, effective tax rate, and basic property attributes. The gap between assessed value and market value is often the most interesting number in real estate.
lat, lng (required)
Assessed value (total, land, improvements), assessment year, annual tax amount, effective tax rate (%), property info (beds, baths, sqft, lot size, year built, last sale)
RentCast API /v1/properties. Keyed (RENTCAST_API_KEY). 50 req/month free tier. Shares API budget with VL-01; cache coordination at the data layer conserves calls.
Weekly. Tax assessments update annually, but 7-day cache aligns with the RentCast call pattern.
A property in Austin, Texas returns an assessed value of $380K with a $4,256 annual tax bill—an effective rate of 1.12%. Stack with vl-01-property-valuation (market AVM of $462K) to see the $82K gap between assessed and market value, which in a non-disclosure state like Texas suggests the owner hasn’t protested their assessment in years. That gap is actionable intelligence for both buyers (what will taxes be after reassessment?) and investors (what’s the true carrying cost?).
vl-01-property-valuation— Market value vs. assessed value gap reveals tax exposurecv-08-tax-rate— Regional tax rate context at the state/county levelvl-08-cost-of-living— Property tax burden relative to area income levels
Surfaces the recent comparable property sales that drive any AVM—sale prices, price per square foot, distance from subject, sale dates, and correlation scores sorted by match quality. Shares the same RentCast API call as VL-01 to conserve the monthly budget.
lat, lng (required)
Array of up to 10 comparable sales (address, sale price, date, sqft, beds/baths, price/sqft, distance, correlation), plus summary statistics (median $/sqft, avg distance, price range)
RentCast API /v1/value-estimate with compCount=10. Keyed (RENTCAST_API_KEY). Same endpoint as VL-01; cache layer prevents duplicate calls.
Weekly. 7-day cache; new comps appear as sales close and record.
A buyer in Bozeman, Montana pulls comps and sees 10 sales ranging from $380K to $575K, with a median of $228/sqft. The top-correlated comp (0.94) sold 0.2 miles away at $245/sqft six weeks ago. The listing they’re evaluating asks $260/sqft. Stack with vl-07-market-trends to see if the market moved 6% in those six weeks, or if the seller is testing a premium. The comps tell you what happened; the trend tells you what’s happening.
vl-01-property-valuation— Comps are the evidence behind the AVM numbervl-07-market-trends— Temporal context for whether comp prices are still currentcv-01-zoning-lookup— Filters comps by zoning match (a comp in C-2 commercial doesn’t inform an R-1 purchase)
Computes a composite insurance risk score from 1 (minimal) to 10 (extreme) by weighting four hazard layers: flood (35%), seismic (25%), wildfire (20%), and wind (20%). Flood and seismic scores come from live government APIs; wildfire and wind use regional estimates pending dedicated data sources.
lat, lng (required)
Composite score (1–10), individual hazard scores with weights and sources, risk label (Minimal through Extreme), hazard availability count
FEMA NFHL Layer 28 (flood, free) + USGS Earthquake Hazards API (seismic, free, M3+ within 200km/30yr) + regional estimates for wildfire and wind (pending USFS and NOAA integration). All hazards queried in parallel.
Monthly. 30-day cache balances seismic activity updates with rate-limit conservation.
A property in Santa Rosa, California returns a composite risk score of 7.4 (Major). Breakdown: flood 1 (Zone X, minimal), seismic 6 (47 M3+ events within 200km), wildfire 8 (regional estimate for Northern California), wind 3 (low). The composite is dominated by wildfire and seismic exposure. Stack with vl-01-property-valuation and vl-06-rental-market to evaluate whether rental yields compensate for the insurance premium this risk profile implies.
vl-02-flood-risk— Provides the flood hazard score inputvl-01-property-valuation— Risk-adjusted valuation when combined with insurance costst-04-natural-hazards— Deeper seismic, volcanic, and tsunami data for coastal properties
Combines HUD Fair Market Rents (the regulatory baseline) with RentCast property-level rent estimates and rental comps. Returns estimated monthly rent, rent range, FMR by bedroom count, and rental comps with distance and correlation. The rent-to-price ratio is computed when paired with VL-01.
lat, lng (required)
Monthly rent estimate with range, HUD FMR by bedroom count (studio through 4BR), rental comps (address, rent, sqft, distance, correlation), rent-to-price ratio (when VL-01 value available)
RentCast API /v1/rent-estimate (keyed, RENTCAST_API_KEY) + HUD FMR API /fmr/statedata/ (keyed, HUD_API_TOKEN, free) + Nominatim. Both degrade gracefully if keys are missing.
Weekly. 7-day cache for rental estimates; HUD FMR updates annually.
An investor evaluating a 3BR rental in Tacoma, Washington gets a RentCast estimate of $2,100/month (range $1,900–$2,300) and a HUD FMR of $1,876 for a 3BR in the area. The market estimate exceeds FMR by 12%, suggesting room above the Section 8 baseline. Stack with vl-01-property-valuation to compute rent-to-price, then with vl-03-tax-assessment and vl-05-insurance-risk to estimate the true net operating income after taxes, insurance, and vacancy.
vl-01-property-valuation— Enables rent-to-price ratio calculation for investment analysisvl-03-tax-assessment— Tax burden relative to rental incomevl-07-market-trends— Rent trends over time to project future cash flow
Tracks real estate market trends at the ZIP-code level—median sale price, average rent, days on market, total listings, and 12-month historical trend data. Computes year-over-year price and rent change percentages. The trend line that tells you whether yesterday’s comps are still today’s market.
lat, lng (required)
Current stats (avg price, avg rent, DOM, listings, YoY price/rent change), 12-month historical trend array (monthly median price, rent, inventory)
RentCast API /v1/market-statistics with historyRange=12. Keyed (RENTCAST_API_KEY). Reverse-geocodes to ZIP via Nominatim.
Weekly. 7-day cache; market stats update as new sales close.
ZIP 97201 (Portland SW) shows median price $548K with a +4.8% YoY change, average DOM of 22 days (down from 31 a year ago), and 142 active listings. The shrinking DOM and rising prices signal an accelerating seller’s market. Stack with vl-04-comparable-sales to see if a specific listing is priced ahead of or behind this trend, and with vl-08-cost-of-living to evaluate whether the price growth is outpacing local income growth.
vl-04-comparable-sales— Trend context for individual comp pricesvl-01-property-valuation— Market momentum behind or against the AVMvl-08-cost-of-living— Affordability context for price trends
Computes a local cost-of-living index relative to the national average by combining BLS Consumer Price Index data (regional price levels) with Census ACS housing costs and median income. Breaks down by category: housing, rent, shelter CPI, overall CPI, and income. A score of 100 means average; above means more expensive.
lat, lng (required)
Composite index (100 = national avg), category breakdown (housing 40%, CPI 30%, income 30%), Census ACS snapshot (median home value, rent, income, rent-as-%-of-income), regional CPI data
BLS CPI API v1 (regional and national series, free, no auth, 25 queries/day) + Census ACS 5-year estimates (state-level, free, optional key) + Nominatim. Three BLS series per call.
Monthly. 30-day cache; CPI updates monthly, ACS annually.
A remote worker comparing Boise, Idaho (composite index 95) to Seattle, Washington (composite index 132) sees that Seattle’s housing index is 168 but its income index is 142, while Boise’s housing index is 102 with an income index of 87. Stack with vl-06-rental-market to compare actual rental costs, and with vl-07-market-trends to see which direction each market is moving. The index tells you where things stand; the trend tells you where they’re going.
vl-06-rental-market— Actual rent figures to ground the indexvl-07-market-trends— Trajectory of costs over timecv-08-tax-rate— Tax burden as part of total cost of livingcv-06-school-district— School quality relative to housing cost for family decisions
grid (the energy lattice)
What powers this place, and how clean, cheap, and reliable is it? Grid covers the energy infrastructure layer: real-time carbon intensity of the local power grid, electricity rates by sector, solar generation potential, power outage detection, fuel-source generation mix, and a composite grid reliability score. It tells you what kind of energy your location runs on, what it costs, and how likely it is to stay on. The family draws primarily on WattTime (carbon), NREL (solar and rates), and the EIA (demand, fuel mix, and retail pricing).
Real-time marginal carbon emissions, electricity rates by sector, solar irradiance and PV system production modeling, hourly grid demand and outage proxy detection, generation mix by fuel type (coal, gas, nuclear, solar, wind, hydro), grid reliability and pricing stability scores.
WattTime API v3 (carbon intensity, free tier index-only, keyed), NREL Utility Rates API v3 (rates, free keyed, 2012 data vintage), NREL Solar Resource + PVWatts V8 (solar potential, free keyed), EIA Open Data API v2 (demand, fuel mix, retail sales; free keyed, ~9K req/hr).
Carbon intensity alone is a number on a dashboard. Stack gr-01-carbon-intensity with gr-05-energy-mix and you can explain why the grid is dirty right now (coal peaking, wind calm). Combine gr-03-solar-potential with gr-02-utility-rate and vl-01-property-valuation, and you’ve built a solar ROI calculator that accounts for both the generation potential and the avoided cost. Layer gr-04-power-outage with gr-06-grid-reliability and vl-05-insurance-risk, and you surface whether a property’s grid infrastructure is a feature or a liability.
Stack gr-03-solar-potential + gr-02-utility-rate + gr-01-carbon-intensity + vl-01-property-valuation for a home in Tucson, Arizona. PVWatts models 7,200 kWh/year from a 4kW system, the residential rate is $0.127/kWh (saving ~$914/year), and the grid runs at MOER 72 (Dirty) during afternoon peaks when solar produces most. The carbon offset is meaningful and the payback period on a $12K system is about 13 years before incentives. Stack with vl-03-tax-assessment to see whether the property assessment would increase with panels, and with gr-05-energy-mix to show the buyer that 40% of their current electricity comes from natural gas.
| ID | Name | Description |
|---|---|---|
| gr-01 | Carbon Intensity | Real-time grid carbon intensity as a 0–100 MOER index via WattTime |
| gr-02 | Utility Rate Lookup | Residential, commercial, and industrial electricity rates via NREL (2012 data vintage) |
| gr-03 | Solar Potential | Solar irradiance and modeled PV system production via NREL Solar Resource + PVWatts V8 |
| gr-04 | Power Outage Map | Grid status via EIA demand anomaly detection (proxy approach—no free real-time outage API) |
| gr-05 | Energy Mix | Hourly electricity generation breakdown by fuel type for the local grid region via EIA |
| gr-06 | Grid Reliability Score | Composite reliability grade from pricing stability trends (proxy for SAIDI/SAIFI) via EIA retail data |
Shows the real-time marginal carbon intensity of your local power grid as a 0–100 MOER index: 0 means the cleanest it gets, 100 means the dirtiest. Updated every 5 minutes. The number that tells you whether charging your EV right now is actually green or just feels green.
lat, lng (required)
MOER index (0–100), grid region identifier, region full name, carbon label (Very Clean through Very Dirty), UTC timestamp
WattTime API v3. Keyed (WATTTIME_USER + WATTTIME_PASS). Free tier: index endpoint for all regions, full MOER values limited to CAISO_NORTH. Bearer token auth with 25-minute proactive refresh.
Every 5 minutes. 300-second cache mirrors the WattTime data update interval.
At 2 PM in Portland, the MOER index reads 34 (Clean) because Bonneville Power is pushing hydro into the grid. By 7 PM, it’s 71 (Dirty) as gas peakers come online for evening demand. Stack with gr-05-energy-mix to see the fuel breakdown behind the index, and with gr-03-solar-potential to understand whether home solar would have been generating during that dirty peak window. The index tells you when to shift load; the mix tells you why.
gr-05-energy-mix— Explains which fuel sources are driving the carbon indexgr-03-solar-potential— Solar generation timing relative to carbon peaksgr-06-grid-reliability— Carbon intensity as one dimension of grid health
Looks up local electricity rates by sector—residential, commercial, and industrial—and identifies the serving utility company. Based on 2012 NREL data; rates have changed since, but the utility service territory and relative pricing structure remain directionally useful.
lat, lng (required)
Utility name, company ID, residential/commercial/industrial rates in $/kWh, sector rate breakdown, data staleness note
NREL Utility Rates API v3. Keyed (NREL_API_KEY). Free, unlimited. Shares 1,000 req/hr limit across all NREL APIs. Data vintage: 2012.
Quarterly. 90-day cache; underlying data is static (2012 vintage).
A small business owner in Denver queries gr-02 and sees that Public Service Company of Colorado serves the area at $0.095/kWh residential, $0.081/kWh commercial. Stack with gr-03-solar-potential to calculate avoided cost from a rooftop PV system, and with gr-01-carbon-intensity to understand whether the utility’s generation mix favors renewables. The 2012 vintage means treating these rates as directional, not definitive.
gr-03-solar-potential— Avoided cost calculation (rate × kWh generated = savings)gr-06-grid-reliability— Rate context feeds into the pricing stability componentvl-08-cost-of-living— Energy cost as a component of overall living expenses
Estimates solar energy potential by combining NREL Solar Resource data (DNI, GHI, tilt irradiance) with PVWatts V8 production modeling. Returns annual and monthly kWh output for a configurable residential PV system, plus capacity factor and the nearest weather station used for TMY data.
lat, lng (required), systemCapacity (optional, default 4 kW)
Annual production (kWh/yr), capacity factor (%), solar radiation (kWh/m²/day), DNI, GHI, tilt irradiance, 12-month production breakdown, system capacity, weather station info
NREL Solar Resource API + NREL PVWatts V8 API. Both free, keyed (NREL_API_KEY). Both queries run in parallel. TMY-based data (static historical). Shares 1,000 req/hr NREL limit.
Annually. 365-day cache; TMY (Typical Meteorological Year) data is inherently historical.
A homeowner in Albuquerque queries gr-03 with the default 4kW system and gets 6,800 kWh/year annual production, a 19.5% capacity factor, and 6.2 kWh/m²/day solar radiation. June produces 720 kWh, December produces 380 kWh. Stack with gr-02-utility-rate at $0.11/kWh to calculate $748/year savings, then with vl-01-property-valuation to evaluate whether the $14K system investment is justified by the 19-year payback (before the 30% federal tax credit brings it under 14).
gr-02-utility-rate— Converts kWh production into dollar savingsgr-01-carbon-intensity— Carbon offset value of solar production during peak dirty hoursvl-01-property-valuation— Solar ROI relative to property value
Tracks grid status using EIA hourly demand data, detecting demand anomalies that may indicate outage events. A proxy approach—no free real-time outage API exists—that flags when demand drops significantly below the 7-day rolling average for a balancing authority region.
lat, lng (required)
Grid status (Normal / Elevated Risk / Potential Disruption), demand deviation percentage, current demand (MWh), rolling average, balancing authority, last 24 hours demand history
EIA Open Data API v2 /electricity/rto/region-data. Keyed (EIA_API_KEY). Free, ~9K req/hr. Maps lat/lng to 7 major balancing authorities (CISO, ERCO, ISNE, NYIS, PJM, MISO, SWPP) via bounding boxes.
Every 15 minutes. 900-second cache; demand data has ~1–2 hour lag.
During a Texas heat wave, gr-04 detects ERCOT demand has dropped 18% below its 7-day average—flagging “Potential Disruption.” This is a proxy signal, not a confirmed outage, but demand drops of this magnitude during peak season correlate with grid stress events. Stack with gr-06-grid-reliability for the longer-term reliability grade, and with gr-05-energy-mix to see if wind generation dropped (a common trigger for ERCOT stress).
gr-06-grid-reliability— Short-term anomaly enriches the long-term reliability assessmentgr-05-energy-mix— Fuel mix context for demand anomaliesvl-05-insurance-risk— Grid reliability as a factor in property risk profiles
Shows the current electricity generation breakdown by fuel type for your local grid region—coal, natural gas, nuclear, solar, wind, hydro, oil, and other. Computed from the latest hourly EIA data, with renewable vs. fossil percentages and color-coded fuel categories.
lat, lng (required)
Fuel sources array (code, name, MWh, percentage, color, renewable flag), total renewable %, total fossil %, total generation (MWh), balancing authority, latest period
EIA Open Data API v2 /electricity/rto/fuel-type-data. Keyed (EIA_API_KEY). Free, ~9K req/hr. Same bounding-box region resolution as GR-04. Fetches 8 rows (one per fuel type).
Hourly. 3,600-second cache; EIA fuel data updates once per hour with ~1–2 hour lag.
At 3 PM in California, CAISO shows: solar 38%, natural gas 28%, imports 15%, nuclear 9%, wind 5%, hydro 4%, other 1%. Renewable share: 47%. Stack with gr-01-carbon-intensity to see how this mix translates to marginal emissions (even at 47% renewable, the marginal unit is often gas), and with gr-03-solar-potential to understand how a home PV system would interact with this grid during its cleanest hours.
gr-01-carbon-intensity— Fuel mix explains the carbon index numbergr-04-power-outage— Fuel diversity context for demand anomaliesgr-06-grid-reliability— Generation diversity as a reliability indicator
Assigns a letter grade (A–F) and 0–100 score to grid reliability using EIA retail pricing data as a proxy. Compares state-level residential rates to the national average, factors in pricing trend stability, and produces a grade that approximates what SAIDI/SAIFI metrics would say if they were freely available via API.
lat, lng (required)
Reliability grade (A–F), reliability score (0–100), state avg residential price (cents/kWh), national avg price, customer count, price trend (stable/increasing/decreasing/volatile), data note explaining proxy methodology
EIA Open Data API v2 /electricity/retail-sales. Keyed (EIA_API_KEY). Free, ~9K req/hr. Fetches 12 months state + national data in parallel. Proxy for SAIDI/SAIFI (EIA Form 861 data requires offline Excel parsing).
Monthly. 30-day cache; retail pricing data updates monthly.
A property in Houston, Texas gets a grade of C (score 58) with a state avg of $0.128/kWh vs. national $0.116/kWh, and a “volatile” price trend reflecting ERCOT’s deregulated market. Compare to a property in Portland, Oregon (grade A, score 89, $0.092/kWh, “stable”). Stack with gr-04-power-outage for short-term grid status, and with vl-05-insurance-risk to factor grid reliability into overall property risk assessment.
gr-01-carbon-intensity— Clean grids tend to correlate with reliabilitygr-04-power-outage— Short-term disruption context for the long-term gradegr-02-utility-rate— Rate data enriches the pricing stability analysisvl-05-insurance-risk— Grid reliability as a risk dimension
cartograph (mapmaker, one who draws the territory)
What does this place look like from above, and how do its layers stack? Cartograph provides the spatial rendering infrastructure for the entire system: base terrain tiles, trail and road networks from OpenStreetMap, water bodies, administrative boundaries, points of interest, sun position calculations, satellite imagery, a layer compositor, and a print-export pipeline. Think of it as the cartographic engine room—most other families produce data about a place, while Cartograph produces the map that data lives on. The first six rills pull from Overpass/OSM and NOAA; the last four are pure computation and configuration.
Base terrain tile configuration (OSM, OpenTopoMap, USGS), trail and road network geometry and metadata, water body features (rivers, lakes, streams), administrative boundary polygons (country through ZIP), points of interest (parks, campgrounds, trailheads, viewpoints), solar position and sun times, satellite imagery tile configuration, multi-layer composition, and print-ready map export metadata.
Overpass API / OpenStreetMap (trails, roads, water, POIs, boundaries; free, no auth), Nominatim (reverse geocoding, free, 1 req/s), OpenTopoMap (terrain tiles, free), USGS National Map (topo tiles, free), NOAA Solar Equations (sun position, inline computation), Mapbox Satellite API (imagery, keyed, 200K free tiles/mo), Hiking Project API (optional trail enrichment, free keyed).
A data rill without a map is a spreadsheet. Stack c-01-base-terrain with c-02-trail-network and t-01-usda-soil-survey, and you can see which trails cross unstable soil. Compose c-04-water-bodies with vl-02-flood-risk and c-05-boundaries, and you can visualize exactly where the flood zone boundary intersects a municipal border and why two adjacent properties have different insurance requirements. The c-09-layer-compositor and c-10-print-cartographer rills are the meta-layer that turns any combination of data rills into a printable, shareable map artifact.
Stack c-01-base-terrain (topo) + c-02-trail-network + c-04-water-bodies + c-06-points-of-interest + c-07-sun-shadow-map for a 40-acre parcel in the Cascade foothills. The topo base shows 200 feet of elevation change, two hiking trails cross the northeast corner, a seasonal creek runs through the western third, and two trailheads sit within 2km. The sun position on the winter solstice shows only 4.2 hours of direct light on the south-facing slope. Feed the stack through c-09-layer-compositor and c-10-print-cartographer to generate a print-ready PDF at 150 DPI that the buyer can bring to the property walk. The map turns data into something you can hold.
| ID | Name | Description |
|---|---|---|
| c-01 | Base Terrain | Tile source configuration for base map layers (OSM, OpenTopoMap, USGS) with no API calls |
| c-02 | Trail Network | Hiking and biking trails from OpenStreetMap with distance, difficulty, and surface metadata |
| c-03 | Road Network | Roads classified by type (motorway through residential) with surface and lane metadata from OSM |
| c-04 | Water Bodies | Rivers, lakes, streams, and wetlands from OSM with geometry type and area calculations |
| c-05 | Boundaries | Administrative boundary polygons (country through ZIP) from OSM + Nominatim with FIPS codes |
| c-06 | Points of Interest | Parks, campgrounds, trailheads, and viewpoints from OSM with amenities and hours metadata |
| c-07 | Sun & Shadow Map | Solar position, sunrise/sunset, golden hour, and twilight via inline NOAA equations (no API calls) |
| c-08 | Satellite Imagery | High-resolution satellite tile configuration via Mapbox (sub-meter, @2x retina tiles) |
| c-09 | Layer Compositor | Meta-rill that orders, blends, and toggles all Cartograph layers into a unified stack |
| c-10 | Print Cartographer | Prepares print-ready map metadata—dimensions, scale bar, legend, attribution (rendering deferred) |
Provides the foundational map tile configuration that every other Cartograph layer renders on top of. Resolves the requested map style (standard, topo, or relief) to a tile URL template with attribution and zoom bounds. Pure computation—no network requests, no API keys, no rate limits.
lat, lng (required), radius_km (optional, default 10), zoom (optional), style (optional: “standard”, “topo”, “relief”; default “topo”)
Tile URL template, attribution, coordinate order, max zoom, computed view center, zoom level, bounding box, layer metadata (z-index, opacity, visibility)
Pure computation. Tile sources: OpenStreetMap (standard, max zoom 19), OpenTopoMap (topo, max zoom 17), USGS National Map (relief, max zoom 16, note: y/x coordinate swap). All tile servers are free and require no auth.
Weekly. 7-day cache; tile configurations are static.
A trail planner sets style: "topo" for a 15km radius around Mt. Hood. The rill returns the OpenTopoMap tile template at zoom 12, bounding box covering the mountain’s south face. Feed this into c-09-layer-compositor as the bottom layer, then stack c-02-trail-network and c-04-water-bodies on top. The base terrain shows contour lines; the overlays show where the trails cross streams.
c-09-layer-compositor— Always the bottom layer in any composed map stackc-08-satellite-imagery— Swappable base layer (terrain vs. satellite)c-02-trail-networkthroughc-07-sun-shadow-map— All data layers render on top of this base
Queries hiking and biking trails from OpenStreetMap via the Overpass API, returning GeoJSON LineStrings with trail name, computed distance (Haversine), surface type, difficulty (mapped from SAC scale), and trail classification. Optionally enriches US trails with elevation gain from the Hiking Project API.
lat, lng (required), radius_km (optional, default 5), trail_type (optional: “hiking”, “biking”, “all”; default “all”)
GeoJSON FeatureCollection (LineStrings), trail metadata array (name, distance, difficulty, surface, type), summary (total trails, total km, type counts), layer metadata
Overpass API / OpenStreetMap (free, no auth, rate limited). OSM tags: highway=path, highway=footway, route=hiking, highway=cycleway, route=bicycle. Optional: Hiking Project API (free keyed, US only, enriches difficulty and elevation).
Weekly. 7-day cache; trail data changes slowly.
A property buyer in the Columbia River Gorge queries trails within 5km and finds 23 hiking trails totaling 87km, including 3 rated “difficult” (SAC mountain_hiking). Stack with c-01-base-terrain (topo) to see the trails against contour lines, then with c-04-water-bodies to identify which trails follow creek corridors, and with c-06-points-of-interest to locate trailheads and viewpoints.
c-01-base-terrain— Trails over topographic contour linesc-04-water-bodies— Trail–creek intersection visibilityc-06-points-of-interest— Trailheads as POI layer on the same mapc-09-layer-compositor— Combined rendering in the multi-layer stack
Queries road and highway data from OpenStreetMap, classifying by type (motorway, trunk, primary, secondary, tertiary, residential) and extracting surface type and lane count. The infrastructure layer that shows how a place connects to everywhere else.
lat, lng (required), radius_km (optional, default 5), road_type (optional: “all”, “highway”, “local”; default “all”)
GeoJSON FeatureCollection (LineStrings), road metadata array (name, classification, surface, lanes), summary (total roads, counts by classification), layer metadata
Overpass API / OpenStreetMap (free, no auth). OSM highway tags: motorway, trunk, primary (highway filter) + secondary, tertiary, residential (local filter).
Weekly. 7-day cache; road networks change slowly.
Evaluating a rural property in eastern Oregon, c-03 shows one secondary road (paved, 2 lanes) connecting to US-26 (primary) 4km north, with no other paved roads within the 5km radius. Stack with ki-03-route-planner for actual drive time to the nearest town, and with c-01-base-terrain to see the road overlay against topography—a critical context for access during winter conditions.
c-01-base-terrain— Roads over terrain for access assessmentki-03-route-planner— Road geometry context for routingc-09-layer-compositor— Road layer in the multi-layer map stack
Queries rivers, lakes, streams, and wetlands from OpenStreetMap, returning Polygon geometry for standing water and LineString for flowing water. Calculates area in km² for lakes and reservoirs. The hydrology layer that tells you where water sits, flows, and could flood.
lat, lng (required), radius_km (optional, default 5), feature_type (optional: “all”, “river”, “lake”, “stream”; default “all”)
GeoJSON FeatureCollection (Polygons for lakes, LineStrings for rivers/streams), water features array (name, type, area for standing water), summary (total features, counts by type), layer metadata
Overpass API / OpenStreetMap (free, no auth). OSM tags: natural=water, water=lake, water=reservoir (standing), waterway=river, waterway=stream (flowing).
Weekly. 7-day cache; water body geometry changes slowly.
A property near Bend, Oregon returns 2 lakes (Mirror Pond 0.042 km², a reservoir 0.18 km²), the Deschutes River, and 5 seasonal streams. Stack with vl-02-flood-risk to see whether FEMA designations align with actual water proximity, and with c-05-boundaries to understand which water features fall within city limits vs. county jurisdiction. Water on a map is scenic; water in a flood zone is a cost.
vl-02-flood-risk— Water proximity vs. FEMA flood zone alignmentc-05-boundaries— Jurisdictional context for water featuresc-02-trail-network— Trail–water crossing visibilityc-09-layer-compositor— Water layer in the multi-layer map stack
Fetches administrative boundary polygons at multiple levels—country, state, county, city, and ZIP code—from OpenStreetMap relations and Nominatim reverse geocoding. Returns MultiPolygon GeoJSON with names and FIPS codes. The most complex rill in the family: sequential Overpass queries, custom QL, and multi-source data fusion.
lat, lng (required), levels (optional, default ["state", "county", "city"])
Boundary results array (level, name, FIPS code, optional GeoJSON MultiPolygon), resolved address components (city, county, state, country, ZIP), layer metadata
Overpass API (custom QL for boundary relations, sequential queries per admin level with rate limiting) + Nominatim reverse geocode (for ZIP and address components, 1 req/s). ZIP comes from Nominatim, not OSM.
Monthly. 30-day cache; administrative boundaries rarely change.
A property straddling the Portland/Beaverton border queries c-05 at ["city", "county"] and gets MultiPolygon boundaries for both cities plus Washington County. Stack with cv-01-zoning-lookup to see whether the zoning changes at the city line, with cv-08-tax-rate to compare tax rates across the boundary, and with c-04-water-bodies to see which jurisdiction controls the creek that runs between the two cities.
cv-01-zoning-lookup— Jurisdictional boundaries that determine zoning authoritycv-08-tax-rate— Tax rate differences across boundary linesc-04-water-bodies— Which jurisdiction controls which waterwayc-09-layer-compositor— Boundary layer in the multi-layer stack
Queries parks, campgrounds, trailheads, and viewpoints from OpenStreetMap. Converts area features (parks as polygons) to centroids so all output is Point geometry. Extracts amenities and opening hours where tagged. The layer that answers “what’s nearby?” for outdoor recreation.
lat, lng (required), radius_km (optional, default 5), poi_type (optional: “all”, “parks”, “camping”, “trailheads”, “viewpoints”; default “all”)
GeoJSON FeatureCollection (Points), POI metadata array (name, type, amenities, opening hours, coordinates), summary (total POIs, counts by type), layer metadata
Overpass API / OpenStreetMap (free, no auth). OSM tags: leisure=park, leisure=nature_reserve (nwr), tourism=camp_site, tourism=caravan_site, highway=trailhead, information=guidepost, tourism=viewpoint (node).
Weekly. 7-day cache; POI data changes slowly in OSM.
Scouting a property in the San Juan Islands, c-06 returns 4 parks (including a 200-acre nature reserve), 2 campgrounds, 6 trailheads, and 3 viewpoints within 5km. Stack with c-02-trail-network to connect trailheads to actual trail geometry, and with c-01-base-terrain to see these points on the topographic map. The POI density relative to population density is a quality-of-life signal that no listing description captures.
c-02-trail-network— Connects trailhead points to trail line geometryc-01-base-terrain— POI dots on the base mapc-07-sun-shadow-map— Sun exposure at viewpoints and campgroundsc-09-layer-compositor— POI layer in the multi-layer stack
Calculates solar position (azimuth and altitude), sunrise, sunset, solar noon, golden hour, and civil twilight for any location and date using inline NOAA solar equations. Zero external dependencies—no API calls, no npm packages, just Julian date math. Phase 2 will add terrain shadow overlays from elevation data.
lat, lng (required), date (optional, default today), time (optional, default “12:00”), include_terrain_shadows (optional, always false in Phase 1)
Solar position (azimuth/altitude in degrees and radians), sun times (sunrise, sunset, solar noon, golden hour, dawn, dusk as ISO strings), shadow overlay (null in Phase 1), layer metadata
Pure computation. NOAA simplified solar equations implemented inline (~200 lines). Julian date conversion, solar mean anomaly, ecliptic longitude, declination, hour angle calculations. No external dependencies.
On demand. Cache TTL 0; computation is instant and date-dependent.
Evaluating garden placement on a property in Corvallis, Oregon on the winter solstice: sunrise 7:44 AM, sunset 4:34 PM (8h 50m daylight), solar altitude at noon is only 21.4°, golden hour starts at 3:42 PM. Stack with fm-03-altitudo to get the terrain profile, and eventually with Phase 2 terrain shadows to see how many of those 8 hours and 50 minutes actually reach the south-facing garden bed.
fm-03-altitudo— Elevation data needed for Phase 2 terrain shadow overlaysgr-03-solar-potential— Sun position context for PV system orientationc-09-layer-compositor— Sun/shadow overlay layer in the map stack
Provides high-resolution satellite imagery tile configuration from Mapbox—sub-meter resolution, @2x retina JPEG tiles, max zoom 22. Returns the tile URL template with access token embedded so downstream rills can overlay aerial photography. Sentinel and Landsat sources are defined but not yet connected.
lat, lng (required), zoom (optional, default 14), source (optional: “mapbox”; “sentinel” and “landsat” defined but not implemented)
Tile URL template with token, imagery metadata (source, resolution, attribution, refresh cadence), view config with bounds, layer metadata
Mapbox Satellite API. Keyed (NEXT_PUBLIC_MAPBOX_TOKEN). Free tier: 200K tile requests/month. @2x retina tiles (512×512 JPEG). Max zoom 22. Attribution: Mapbox, OpenStreetMap, Maxar.
Monthly. 30-day cache; satellite tile configs are stable. Imagery itself is a continuously updated composite.
Inspecting a rural property before visiting in person: c-08 provides the satellite base layer at zoom 16, revealing tree cover density, building footprints, and access road conditions that the topo base doesn’t show. Stack with c-05-boundaries to overlay parcel lines on the satellite view, and with c-02-trail-network to see whether the trail visible in the imagery is formally mapped in OSM.
c-01-base-terrain— Swappable: satellite vs. topo as the base layerc-05-boundaries— Parcel and jurisdictional lines over aerial photographyc-09-layer-compositor— Satellite as the base layer in the compositor stack
Meta-rill that consumes layer metadata from C-01 through C-08 and composes them into an ordered layer stack. Manages z-ordering, opacity, blend modes, and visibility toggling. Pure client-side computation that turns individual data layers into a unified, interactive map. The conductor of the cartographic orchestra.
layers (required, LayerMeta array from C-01–C-08), options (optional: z-order array, opacity map, blend modes, hidden layers)
Composite config (ordered layer stack with resolved z-index, opacity, blend modes), active/visible layers array, combined attribution, union bounding box, self layer metadata
None. Pure client-side computation. Consumes outputs from all 8 Cartograph data rills. Source type: compound.
On demand. Cache TTL 0; recomputed whenever input layers change.
Building a property scouting map: feed c-01 (topo base, z:0), c-04 (water, z:10), c-02 (trails, z:20), c-05 (boundaries, z:30, opacity 0.5), and c-06 (POIs, z:40) into the compositor. It resolves the z-order, deduplicates attributions, computes the union bounding box, and outputs a single config that a map renderer can consume. Toggle boundaries off, and the compositor recalculates the active layer set without refetching any data.
c-01throughc-08— All Cartograph data rills feed into the compositorc-10-print-cartographer— Compositor output becomes print input
Takes the composite layer stack from C-09 and prepares print-ready map metadata: pixel dimensions at configurable DPI, a “nice” scale bar, auto-generated legend entries, combined attribution text, and estimated file size. The actual PDF/PNG rendering is deferred—this rill generates everything a renderer would need to produce the artifact.
composite_config (required, from C-09), print_options (optional: format pdf/png, size a4/letter/custom, DPI default 150, legend toggle, title default “Slipstream Map”)
Print config (dimensions in px and mm, DPI, scale bar with length in km and px, legend entries, title, attribution, north arrow flag), print metadata (pixel dimensions, scale string, estimated file size in KB)
None. Pure computation consuming C-09 output. Source type: compound. Scale bar candidates: 0.1, 0.25, 0.5, 1, 2, 5, 10, 20, 50, 100, 200, 500 km.
On demand. Generated per export request.
After compositing a 5-layer map of a property in the Cascades, the buyer exports at A4/150 DPI. c-10 returns: 1240×1754 px, scale bar “5 km” (248 px long), 5 legend entries auto-generated from visible layer IDs, combined attribution “OpenTopoMap CC-BY-SA | OpenStreetMap contributors | NOAA Solar Equations,” and estimated file size 1.2 MB (PDF). The buyer prints it and walks the property with data in hand.
c-09-layer-compositor— The only input; compositor output is print input
forma (shape, form, figure)
What is the shape of this place, and what can you compute from that shape? Forma provides the geometric and spatial computation primitives: forward and reverse geocoding, spatial operations (buffer, intersect, union, difference), elevation profiling, viewshed analysis, building footprint extraction, and parcel boundary lookup. Where Cartograph renders maps, Forma computes on geometry. It’s the math layer beneath the visual layer—the family that answers questions like “what can I see from this hilltop?” and “how big is this parcel?”
Forward and reverse geocoding, GeoJSON spatial operations (buffer, intersect, union, difference), single-point and profile elevation from DEM data, simplified viewshed analysis, building footprint polygons, and parcel/cadastral boundary lookup with ownership and use codes.
Mapbox Geocoding v6 (geocoding, keyed, 100K free/mo), Turf.js (spatial ops, client-side, free), OpenTopoData (elevation, free public instance, 1K req/day), Overture Maps (building footprints, free open data on S3—pipeline not yet wired), Regrid Parcel API (cadastral, paid after trial).
An elevation number is data. Stack fm-03-altitudo with c-07-sun-shadow-map and gr-03-solar-potential, and it becomes an answer to “will a south-facing slope at this elevation produce enough solar energy to justify panels?” Compose fm-02-spatium (buffer 500m around a school) with cv-01-zoning-lookup and cv-10-legal-compliance, and you can identify which parcels within the buffer zone have active code violations—geometry as a query filter. Combine fm-06-fundus with vl-01-property-valuation and cv-04-property-regulations, and you surface whether a parcel’s acreage supports subdivision under current zoning.
Stack fm-01-geocodex (resolve address to coordinates) + fm-03-altitudo (elevation profile across the parcel) + fm-04-prospectus (viewshed from the building site) + fm-06-fundus (parcel boundaries and acreage) + cv-01-zoning-lookup (permitted uses). A developer evaluating a 2.3-acre parcel in the Columbia Gorge gets: elevation ranges from 320m to 368m across the site (15% grade on the north slope), the viewshed shows 73% visibility from the proposed building pad, the parcel is zoned RR (Rural Residential) allowing one dwelling per acre, and the owner is a family trust. The composition surfaces that the site can support two dwellings with views, but the north slope will require engineered foundations.
| ID | Name | Description |
|---|---|---|
| fm-01 | Geocodex | Forward and reverse geocoding via Mapbox v6 with confidence scoring and structured address |
| fm-02 | Spatium | GeoJSON spatial operations (buffer, intersect, union, difference) via Turf.js, zero API overhead |
| fm-03 | Altitudo | Point and profile elevation from OpenTopoData with multiple DEM datasets (SRTM 90m/30m, NED 10m) |
| fm-04 | Prospectus | Simplified viewshed analysis from an observer point using sampled terrain elevation data |
| fm-05 | Vestigium | Building footprint extraction from Overture Maps (stub—data pipeline not yet wired) |
| fm-06 | Fundus | Parcel boundaries, ownership, use codes, zoning, and acreage via Regrid (US/Canada, paid) |
Converts addresses to coordinates (forward) and coordinates to structured addresses (reverse) via the Mapbox Geocoding v6 API. Returns confidence scoring and position accuracy (rooftop, parcel, interpolated, approximate). The translation layer between human addresses and machine coordinates.
query (forward) or lat, lng (reverse). Provide one or the other.
Coordinates [lng, lat], structured address (full, name, postcode, place, region, country), confidence (exact/high/medium/low), accuracy (rooftop/parcel/point/interpolated/approximate), feature type
Mapbox Geocoding v6. Keyed (MAPBOX_ACCESS_TOKEN). Free tier: 100K req/month. 1,000 req/min. Fetches limit=1 (single best result).
Monthly. 30-day cache; addresses change slowly.
Forward-geocode “3847 SE Hawthorne Blvd, Portland, OR” and get rooftop-accuracy coordinates at [−122.6312, 45.5118] with “exact” confidence. Feed those coordinates into any lat/lng-based rill in the system—cv-01-zoning-lookup, vl-01-property-valuation, t-01-usda-soil-survey—and the address becomes a key that unlocks the entire data layer. Geocodex is the front door.
fm-02-spatium— Geocoded point as input to spatial operationsfm-06-fundus— Address resolution before parcel lookup- Any lat/lng rill — Geocodex is the address-to-coordinates translation for the entire system
Performs spatial analysis operations on GeoJSON geometries: buffer a point or polygon, compute the intersection, union, or difference of two geometries, and measure the resulting area. All computation runs client-side via Turf.js—zero API calls, zero latency, zero cost. Includes self-intersection validation via turf.kinks().
geometry (required, GeoJSON Feature), operation (required: buffer/intersect/union/difference), params (optional: radius, units, secondGeometry)
Result GeoJSON Feature (or null if empty, e.g., non-overlapping intersect), area in m², operation type, validity flag
Turf.js (client-side JavaScript library). No external API calls. Free, MIT-licensed, unlimited. Dynamically imported for code-splitting.
On demand. Cache TTL 0; pure computation.
Buffer a school location by 500 meters, then intersect that buffer with cv-01-zoning-lookup results to find which zoning designations fall within the buffer. Or take two fm-06-fundus parcel polygons and compute their union to model a lot merger, then check whether the merged area meets the minimum lot size from cv-04-property-regulations. Geometry becomes a query language.
fm-01-geocodex— Geocoded points as buffer centersfm-06-fundus— Parcel polygons as inputs to spatial operationscv-01-zoning-lookup— Spatial filtering of zoning data
Returns elevation above sea level for a single point or computes an elevation profile along a path, using OpenTopoData with configurable DEM datasets (SRTM 90m global, SRTM 30m, NED 10m for US). Profile mode batches path queries at 100-point boundaries with rate-limit delays and computes cumulative Haversine distance.
lat, lng (required), path (optional, array of [lat, lng] pairs for profile), dataset (optional: “srtm90m”, “srtm30m”, “ned10m”; default “srtm90m”)
Elevation in meters, dataset used, coordinates, optional profile array (lat, lng, elevation, cumulative distance in meters)
OpenTopoData public instance. Free, no auth. Rate limited: 1 req/sec, 1K req/day, max 100 locations per request. Pipe-separated coordinate format (lat,lng order).
Annually. 365-day cache; DEM data is static and never changes.
Profile a 2km driveway path to a rural property: 40 sample points reveal 85 meters of elevation gain with a maximum grade of 18% near the top. Stack with c-03-road-network to see whether an existing road follows the same path, and with c-07-sun-shadow-map to understand sun exposure along the slope. The elevation profile turns “steep driveway” into “18% grade for 200 meters starting at the second switchback.”
fm-04-prospectus— Elevation data feeds viewshed analysisc-07-sun-shadow-map— Elevation context for sun exposure calculationsgr-03-solar-potential— Site elevation and slope affect solar panel performance
Computes a simplified viewshed from an observer point: samples 40 terrain points across 8 compass bearings at 5 distance rings, then calculates what percentage of the surrounding terrain is visible. A working approximation—the full Bresenham ray-casting algorithm with intermediate obstruction checking is planned for Phase 2.
lat, lng (required), observerHeight (optional, default 1.7m), radius (optional, default 1000m)
Observer elevation (m), observer height, analysis radius, visibility percentage (0–100%), sample count, visible count
OpenTopoData (SRTM 90m, hardcoded). Free, no auth. Two API calls total: 1 for observer elevation, 1 for 40 sample points. 1.1-second rate-limit delay between calls.
Annually. 365-day cache; terrain is static.
Evaluating a building site on a ridge in the Willamette Valley: fm-04 returns 73% visibility from a 1.7m observer height within a 1km radius. Stack with c-07-sun-shadow-map to understand sun angles across the visible terrain, and with c-08-satellite-imagery to see what the visible terrain actually looks like. A 73% viewshed on a ridge suggests the remaining 27% is blocked by the ridge itself to the east—a constraint on morning light but not on western views.
fm-03-altitudo— Elevation data is the input to viewshed computationc-07-sun-shadow-map— Sun angles across visible terrainc-08-satellite-imagery— Visual ground truth for the viewshed
Designed to extract building footprint polygons from Overture Maps’ 2.3 billion worldwide buildings dataset. Currently a stub—computes the bounding box and summary statistics framework, but returns an empty buildings array because the GeoParquet data pipeline (DuckDB WASM or serverless function) is not yet wired. The shape of things to come.
lat, lng (required), radius (optional, default 500m)
Buildings array (currently empty), summary (total count, with-height count, with-floors count, class distribution), bounding box, center point
Overture Maps (open data on S3, GeoParquet format). Free, no auth, no rate limits. 2.3B buildings worldwide. Pipeline not connected—would require DuckDB WASM or pre-processed GeoJSON.
Quarterly. 90-day cache; Overture releases are semi-annual.
Once the data pipeline is connected, query building footprints within 500m of a property to understand neighborhood density, building heights, and structure classes. Stack with c-08-satellite-imagery to cross-reference footprints against aerial photography, and with fm-06-fundus to see how building footprints relate to parcel boundaries. A building that extends beyond its parcel boundary is either an encroachment or a shared-wall structure—both worth knowing before purchase.
fm-06-fundus— Building footprints relative to parcel boundariesc-08-satellite-imagery— Footprint verification against aerial imageryc-09-layer-compositor— Building footprint layer on the map stack
Returns parcel boundaries, ownership information, land use codes, zoning, and acreage for US and Canadian locations via the Regrid Parcel API. The cadastral layer that answers the fundamental real estate question: where does this property begin and end, who owns it, and what is it used for?
lat, lng (required)
Parcel GeoJSON Feature (Polygon with properties), owner name, use description, acreage, address, parcel number, source agent (county)
Regrid Parcel API v2. Keyed (REGRID_API_TOKEN). Paid (~$49/month after 30-day trial). Billed per parcel record, not per API call. US and Canada only (150M+ parcels). limit=1 to minimize billing.
Quarterly. 90-day cache; county assessor data updates quarterly to annually.
Query a property in Detroit and get: parcel polygon, owner “Wayne County Land Bank Authority,” use code “Residential Vacant Land,” acreage 0.14, parcel number 182127, source agent “Wayne County.” Stack with cv-01-zoning-lookup to check development potential, with vl-01-property-valuation to get the AVM, and with fm-02-spatium to compute whether two adjacent Land Bank parcels could be merged into a buildable lot.
fm-01-geocodex— Address resolution before parcel lookupfm-02-spatium— Parcel polygons as inputs to spatial operations (merge, intersect)cv-01-zoning-lookup— Parcel boundaries aligned with zoning designationsvl-01-property-valuation— Property value for the parcel
kinesis (movement, motion)
How do people move through this place, and what patterns emerge from that movement? Kinesis is the personal mobility layer: live location tracking with consent management, GPS route recording, A-to-B route planning, activity import from Strava and GPX/KML files, commute pattern analysis, geofence monitoring, heatmap generation, elevation profiling, distance calculation, and place history clustering. Unlike the other Polis families that describe a place’s properties, Kinesis describes how a person moves through it. Most rills are pure local computation over stored track data—no API calls, no keys, no cost.
Live GPS position with consent management, GPS track recording and GeoJSON export, turn-by-turn routing, Strava/GPX/KML activity import, commute pattern detection, geofence zone monitoring with enter/exit/dwell events, activity density heatmaps, elevation profile computation, Haversine and routed distance, and DBSCAN-based place clustering with visit history.
Browser Geolocation API (live position, free, no key), Mapbox Directions API (routing, keyed, 100K free/mo) with OSRM fallback (free, non-commercial), Strava API v3 (activity import, OAuth 2.0, 200 req/15min), DOMParser (GPX/KML file parsing, client-side). 7 of 10 rills are pure local computation with zero external dependencies.
A GPS track by itself is a line on a map. Stack ki-02-route-recorder with ki-08-elevation-profile and c-01-base-terrain, and it becomes a topographic trail report. Compose ki-10-place-history with ki-05-commute-analyzer and c-05-boundaries, and you can see which jurisdictions a person spends time in and how their commute patterns cross municipal lines. Layer ki-07-heatmap-generator with c-06-points-of-interest and c-02-trail-network, and you surface which trails and parks actually get used versus which ones just exist on a map. Movement data turns static geography into lived experience.
Stack ki-04-activity-importer (pull 6 months of Strava runs) + ki-07-heatmap-generator (density map of all runs) + ki-08-elevation-profile (per-run elevation stats) + ki-10-place-history (cluster frequent run start/end locations) + c-02-trail-network (overlay OSM trail geometry). The heatmap reveals that 72% of running activity concentrates on 3 trail corridors in Forest Park. Place history clusters identify “home” (nighttime start point), “work” (weekday lunch runs), and “trailhead” (weekend morning starts). Elevation profiles show a preference for routes with 200–400m total ascent. The composition transforms raw GPS logs into a legible story about how someone uses their landscape.
| ID | Name | Description |
|---|---|---|
| ki-01 | Live Location Tracker | Browser Geolocation with full consent management (grant/deny/revoke/scope) |
| ki-02 | Route Recorder | GPS track recording with start/stop/pause, waypoint capture, and GeoJSON export |
| ki-03 | Route Planner | A-to-B directions via Mapbox with driving/walking/cycling profiles and OSRM fallback |
| ki-04 | Activity Importer | Imports from Strava (OAuth 2.0) and GPX/KML files, normalizing to KinesisTrack format |
| ki-05 | Commute Analyzer | Detects commute patterns from track history—frequent routes, day-of-week timing, best departure |
| ki-06 | Geofence Manager | Circle and polygon geofences with enter/exit/dwell event detection (turf.js math inlined) |
| ki-07 | Heatmap Generator | Activity density grid from historical tracks with top-10 hotspot identification |
| ki-08 | Elevation Profile | Altitude profile and grade computation for GPS tracks with 3D coordinates |
| ki-09 | Distance Calculator | Haversine great-circle distance and bearing between two points (routed distance planned) |
| ki-10 | Place History | DBSCAN clustering of visited locations with visit counts, dwell time, and place categorization |
Wraps the Browser Geolocation API with a full consent management layer: grant, deny, revoke, scope (foreground/background), prompt counting, and localStorage persistence. Returns position with accuracy, altitude, speed, and heading. The privacy-first foundation that every other Kinesis rill builds on.
None required. Location consent must be granted before position data flows.
Latitude, longitude, accuracy (m), altitude (m, optional), speed (m/s, optional), heading (degrees, optional), timestamp, consent state
Browser Geolocation API (navigator.geolocation). Free, built into all browsers, no API key. getCurrentPosition (one-shot) and watchPosition (continuous) with configurable accuracy and timeout.
Real-time. Cache TTL 0; position is inherently ephemeral.
A user visiting a property site grants foreground location consent. ki-01 begins streaming position updates, feeding ki-02-route-recorder to capture the property walk as a GPS track. The consent layer ensures the user explicitly opted in, tracks how many times they’ve been prompted, and provides a clean revoke path. Stack with ki-06-geofence-manager to get an alert when they enter or leave the parcel boundary from fm-06-fundus.
ki-02-route-recorder— Live position stream feeds track recordingki-06-geofence-manager— Position updates trigger geofence checkski-10-place-history— Current location as a visit data point
Records GPS tracks with start, stop, and pause controls. Builds a KinesisTrack from live waypoints: computes total distance (Haversine), elevation gain/loss, duration accounting for pauses, transport mode inference from speed, and exports to GeoJSON FeatureCollection with coordinate timestamps.
recordingAction (optional: “start”, “stop”, “pause”, “idle”)
KinesisTrack (waypoints, GeoJSON LineString, total distance, elevation gain/loss, duration, inferred transport mode), recording state, waypoint count, elapsed time
Browser Geolocation API via KI-01. Free, no key. Transport mode inferred from average speed: walk <1.7 m/s, run 1.7–5.5, cycle 5.5–8.3, drive >8.3.
Real-time during recording. Cache TTL 0.
Record a property boundary walk: start recording at the front gate, walk the perimeter, stop at the gate. The track captures 847 waypoints over 23 minutes, computes 1.4km distance, 42m elevation gain, and infers “walk” mode. Export to GeoJSON and stack with c-01-base-terrain to overlay the walk on a topo map, or with fm-06-fundus to compare the walked perimeter against the official parcel boundary.
ki-08-elevation-profile— Elevation analysis of the recorded trackki-05-commute-analyzer— Recorded tracks feed commute pattern detectionki-07-heatmap-generator— Recorded tracks contribute to the activity density mapc-01-base-terrain— Track overlay on the base map
Computes A-to-B directions with driving, walking, and cycling profiles via the Mapbox Directions API. Returns route geometry, total distance, duration, and turn-by-turn navigation steps with maneuver types. Falls back automatically to the OSRM demo server when Mapbox is unavailable.
origin {lat, lng} (required), destination {lat, lng} (required), profile (optional: driving/walking/cycling, default driving)
Route GeoJSON LineString, total distance (m), estimated duration (s), turn-by-turn steps array (instruction, distance, duration, road name, maneuver type/modifier/location)
Primary: Mapbox Directions API v5. Keyed (NEXT_PUBLIC_MAPBOX_TOKEN). 100K free/month, $2/1K after. 300 req/min (rate-limited in code at 200ms intervals). Fallback: OSRM demo server (free, non-commercial, 1 req/s).
Daily. 24-hour cache for route geometry; road networks don’t change hourly.
How far is a rural property from the nearest town? Route from the property gate to the Safeway in Hood River: 28.4km, 34 minutes driving via OR-35. The turn-by-turn shows 2km of gravel road before reaching the highway. Stack with c-03-road-network to see the road classifications along the route, and with ki-05-commute-analyzer after recording a few actual drives to see how the real commute time compares to the modeled time.
ki-05-commute-analyzer— Planned route vs. actual commute patternsc-03-road-network— Road classification context for the routeki-09-distance-calculator— Crow-flies distance vs. routed distance comparison
Imports fitness activities from Strava via OAuth 2.0 and from GPX/KML file uploads, normalizing everything to the KinesisTrack format. Parses Strava streams (latlng, altitude, distance, time), GPX trackpoints with elevation and timestamps, and KML coordinate triplets. Handles Strava rate limits with user feedback.
platform (optional: “strava” or “file”), file (optional: GPX or KML file object)
Array of imported KinesisTracks, import count, total distance across all imports. Each track includes full metadata: distance, elevation, duration, transport mode, GeoJSON geometry.
Strava API v3 (OAuth 2.0, STRAVA_CLIENT_ID, 200 req/15min, 2K/day). OAuth token exchange handled externally; this rill assumes a valid access token. GPX/KML parsed via DOMParser (client-side, free). KML coordinates are already [lng, lat, alt] order.
On demand. Cache TTL 0; imports are user-initiated.
Import 6 months of Strava running data (142 activities) and 3 GPX files from a handheld GPS. All 145 tracks normalize to KinesisTrack format. Feed into ki-07-heatmap-generator for density analysis, ki-05-commute-analyzer for pattern detection, ki-10-place-history for frequently visited locations, and ki-08-elevation-profile for per-run elevation stats. The importer is the data on-ramp; the analytical rills are the value.
ki-05-commute-analyzer— Imported tracks feed commute pattern detectionki-07-heatmap-generator— All imported tracks contribute to the density mapki-08-elevation-profile— Per-track elevation analysis from imported altitude dataki-10-place-history— Start/end points of imported tracks feed place clustering
Analyzes commute patterns from recorded and imported tracks. Clusters routes by start/end proximity (500m threshold), detects frequently traveled corridors, computes average travel time by day-of-week and hour-of-day, identifies the optimal departure time, and infers primary transport mode. Pure local computation on stored track history.
timeRange (optional: start/end ISO dates for filtering)
Commute patterns array (route clusters with occurrence count, avg duration, avg distance, day-of-week/hour distribution, transport mode), overall avg duration/distance, best departure time, departure buckets, total trips
Pure local computation over stored KinesisTrack data. No external API calls. Free, no key, no rate limits. Greedy clustering: assigns each track to the first cluster with matching start/end points or creates a new one.
Hourly. 1-hour cache; recomputed as new tracks are added.
After 3 months of recording commutes between home and office, ki-05 identifies 2 route clusters: I-84 (22 min avg, 47 trips) and Surface Streets (28 min avg, 12 trips). Best departure: 7:15 AM (18 min avg vs. 26 min at 8:30 AM). Stack with c-03-road-network to visualize the two routes, and with ki-08-elevation-profile to compare elevation changes (relevant if considering a cycling commute alternative).
ki-03-route-planner— Planned vs. actual commute comparisonc-03-road-network— Road classification along commute corridorski-07-heatmap-generator— Commute corridor density visualization
Defines geofence zones as circles or polygons and monitors position for enter, exit, and dwell events. Circle fences use Haversine distance; polygon fences use ray-casting point-in-polygon. All spatial math is inlined (turf.js equivalent) to minimize bundle size. Tracks active alerts and time spent inside each zone.
geofences (optional: array of fence definitions with center/radius or polygon coordinates), currentPosition (optional: {lat, lng})
All defined geofences, active alerts (enter/exit/dwell events with timestamps), array of zone IDs the user is currently inside
Pure local computation. Haversine distance, ray-casting point-in-polygon, and circle-to-polygon (64-vertex approximation) implemented inline. No external API calls. Free, no key.
Real-time. Cache TTL 0; checked on each position update.
Define a geofence around a property parcel polygon from fm-06-fundus with alertOnEnter: true and dwellThreshold: 300 (5 minutes). When the user walks onto the property during a site visit, ki-06 fires an enter event. After 5 minutes on-site, it fires a dwell event—useful for logging how long was spent on each property in a multi-site tour. Stack with ki-01-live-location for the position stream.
ki-01-live-location— Position updates trigger geofence checksfm-06-fundus— Parcel polygon as a geofence boundaryc-05-boundaries— Administrative boundaries as geofence zones
Generates an activity density heatmap from historical GPS tracks. Aggregates all coordinate points into a grid (configurable resolution, default ~111m cells), computes frequency per cell, normalizes weights 0–1, and outputs GeoJSON Points suitable for MapLibre GL heatmap layer rendering. Identifies top 10 hotspots by visit density.
timeRange (optional: start/end ISO dates), resolution (optional: grid cell size in degrees, default 0.001)
GeoJSON FeatureCollection (Points with weight property), top 10 hotspots (lat, lng, count, rank), total points aggregated, grid cell count
Pure local computation over stored KinesisTrack data. No external API calls. Free, no key. Includes a demo data generator with 5 San Francisco landmark clusters for testing without real data.
Hourly. 1-hour cache; recomputed as new tracks are added.
After importing 6 months of Strava running data, ki-07 generates a heatmap showing activity concentrated along 3 trail corridors in Forest Park, with the hottest cell at the Wildwood Trail junction (342 coordinate hits). Stack with c-02-trail-network to overlay the heatmap on formal trail geometry—the alignment (or misalignment) between official trails and actual usage patterns is the signal.
c-02-trail-network— Activity density relative to formal trail geometryc-06-points-of-interest— Hotspot proximity to parks and trailheadsc-01-base-terrain— Heatmap overlay on the base map
Computes altitude profile and elevation statistics for a GPS track with 3D coordinates: total ascent, total descent, max and min elevation, grade percentage per segment, average and maximum grade. Suitable for rendering as an SVG area chart showing elevation along the route distance.
track (required: KinesisTrack with altitude data in coordinates)
Elevation points array (distance, elevation, grade, lat, lng), total ascent/descent (m), max/min elevation (m), avg grade (%), max grade (%), total distance (m)
Pure local computation on 3D coordinate arrays. Altitude comes from GPS sensor, Strava streams, or GPX file elevation elements. Free, no key. Includes a demo profile generator (5km hilly route with sine-wave elevation).
On demand. Cache TTL 0; computed per track.
Analyze a recorded property walk: 1.4km distance, 42m total ascent, 38m total descent, max elevation 312m, min 270m, average grade 6.2%, max grade 22% at the 800m mark. Stack with c-01-base-terrain (topo) to correlate the grade spikes with visible contour line density, and with fm-03-altitudo for DEM-validated elevation (GPS altitude can drift 5–15m from barometric pressure changes).
fm-03-altitudo— DEM-validated elevation to cross-check GPS altitudec-01-base-terrain— Contour line context for grade analysiski-02-route-recorder— Source of tracked routes to profile
Calculates the great-circle distance between two geographic points using the Haversine formula and computes the initial bearing (compass direction). Returns distance in meters with unit conversion utilities and a 16-point compass direction. Routed (road-following) distance is planned but currently returns null.
from {lat, lng} (required), to {lat, lng} (required), method (optional: “haversine”; “routed” defined but returns null)
Haversine distance (m), routed distance (null, planned), unit, bearing (degrees), compass direction (16-point: N, NNE, NE, ENE, etc.)
Pure Haversine computation (Earth radius 6,371,008.8m). Free, no key. Routed distance would use Mapbox Directions API via KI-03 (not yet wired).
On demand. Cache TTL 0; pure computation.
How far is the property from the nearest fire station? Haversine says 3.2km bearing SSW. Stack with ki-03-route-planner for the actual driving distance (likely 5–8km on roads), and with vl-05-insurance-risk because fire station proximity affects homeowner’s insurance premiums. The crow-flies distance is a quick filter; the routed distance is the reality check.
ki-03-route-planner— Haversine vs. routed distance comparisonvl-05-insurance-risk— Distance to emergency services as a risk factorc-06-points-of-interest— Distance to specific POIs
Clusters visited locations from GPS track history into named places using DBSCAN (Density-Based Spatial Clustering of Applications with Noise). Detects frequently visited spots, computes visit counts and dwell times, and categorizes places by visit patterns: nighttime visits suggest “home,” weekday daytime suggests “work,” weekend visits suggest “leisure.”
timeRange (optional: start/end ISO dates), clusterRadius (optional: clustering radius in meters, default 100)
Places array (centroid, first/last visit, unique visit count, total dwell time, inferred category), total visits, top 10 most-visited places sorted by frequency
Pure local computation. Full DBSCAN implementation with Haversine distance metric. Extracts start and end coordinates from each track as visit points. Min 2 points per cluster. Free, no key. Includes a demo generator with simulated home/work/gym patterns.
Hourly. 1-hour cache; recomputed as new tracks are added.
After 3 months of activity tracking, ki-10 identifies 12 place clusters. The top 3: “home” (48 visits, 87% nighttime, category: home), “office” (41 visits, 92% weekday daytime, category: work), and “Forest Park trailhead” (18 visits, 78% weekend morning, category: leisure). Stack with c-05-boundaries to see which jurisdictions contain the detected places, and with ki-05-commute-analyzer to connect the place clusters with the routes between them.
ki-05-commute-analyzer— Place clusters as commute route endpointsc-05-boundaries— Jurisdictional context for detected placeski-07-heatmap-generator— Place clusters overlaid on the activity heatmapc-06-points-of-interest— Detected places correlated with formal POIs
Communitas — The Human World
Once you’ve mapped the geology, the ecology, and the regulatory grid, the question shifts: who actually lives here? Communitas is the human and social layer—neighborhoods, demographics, community knowledge, gathering places, live events, personal health, environmental health, and the interpretive framework of the stars. This is the domain where data becomes personal, where “place” becomes “home.”
These rills draw from Census data, community contributions, signed observations, wearable devices, USDA food databases, EPA environmental health records, Swiss Ephemeris computations, and the living knowledge that neighbors carry but no federal database records. Individually, a census tract is a statistical boundary. Stacked with community observations, health data, and event calendars, it becomes a portrait of what it actually feels like to live somewhere.
communis (shared, common, public)
What do the people who live here actually know? Communis is the community intelligence layer—the observations, local knowledge, trail conditions, water reports, noise measurements, and mutual aid networks that exist only because people contribute them. No satellite can see this. No federal database records it. This is the knowledge layer that makes Slipstream alive: geotagged, timestamped, signed, and reputation-scored observations from the people who actually walk these trails, drink this water, and listen to these neighborhoods at 2 AM.
Community observations (wildlife, conditions, events), foraging commons with governed privacy, trail conditions, noise measurements, water reports, mutual aid networks, local ecological knowledge, neighborhood bulletins, citizen sensor data (PurpleAir, weather stations), and harvest reports.
User contributions with cryptographic provenance (signed, timestamped, append-only log). Cross-referenced with institutional data: USGS gauges for water, NWS for weather correlation, BTS noise maps for baseline. PurpleAir API (citizen air quality, free). Community-governed privacy settings determine location precision.
A USGS stream gauge tells you the river’s level. cm-05 tells you what the water actually looks like where no gauge exists—algae blooms, clarity, flooding. Stack cm-01 (observations) with cm-05 (water truth) and vi-03 (toxic release), and a community can document environmental contamination with timestamped evidence, regulatory context, and demographic equity framing. Layer cm-07 (local ecological knowledge) with fl-06 (phenology) and historical climate data, and you surface a documented narrative of ecological change that bridges oral tradition and quantitative measurement.
Stack cm-01 + cm-05 + cm-09 + vi-03 + vi-05 for a neighborhood in Portland’s Cully district. Community water observations flag discoloration in a creek that flows through a park. Citizen sensors show air quality readings significantly worse than the nearest EPA monitor. The toxic release rill shows a TRI facility upwind. Environmental justice screening confirms elevated pollution burden overlapping with vulnerable demographics. This surfaces a signal that the official monitoring network may underrepresent exposure in this community—a finding visible only when community-sourced and institutional data are stacked together.
| ID | Name | Description |
|---|---|---|
| cm-01 | Community Observations | General-purpose geotagged observation feed with signed provenance and reputation scoring |
| cm-02 | Foraging Commons | Community-governed foraging intelligence with privacy-graded location precision |
| cm-03 | Trail Truth | Real-time trail conditions from community reports cross-referenced with weather |
| cm-04 | Noise Witness | Community noise measurements from phone microphones with location and source tagging |
| cm-05 | Water Truth | Community water condition reports cross-referenced with USGS gauge data |
| cm-06 | Mutual Aid Network | Tool libraries, seed exchanges, skill shares, and repair cafes with community governance |
| cm-07 | Local Ecological Knowledge | Digitized oral tradition from experienced locals with attribution and provenance |
| cm-08 | Neighborhood Pulse | Hyperlocal community bulletin with structured event-type tagging |
| cm-09 | Citizen Sensor Network | Aggregation of community-deployed PurpleAir, weather stations, and soil probes |
| cm-10 | Harvest Reports | Ground-truth foraging outcomes that feed back into habitat prediction models |
General-purpose observation feed where anyone can pin a geotagged note—wildlife sighting, trail condition, water level, noise event, construction activity, anything worth knowing about a place. Think local iNaturalist for everything, not just species.
lat, lng (required); observation_type, text, media (optional); contributor’s signed identity
Observation feed for area: geotagged entries with confidence scores, contributor reputation, corroboration count, and cross-reference status
User contributions with cryptographic provenance. Cross-referenced with institutional APIs (USGS, NWS) where overlap exists. Free.
Real-time. Observations appear immediately; confidence scores update as cross-references and corroborations arrive.
A resident near Tryon Creek State Park in SW Portland pins an observation: “Large tree down across North Creek Trail, blocks full trail width, happened overnight.” Two other hikers corroborate within hours. The confidence score climbs. Cross-referenced with last night’s NWS wind advisory (sustained 45 mph), the observation gains institutional backing. Trail Truth (cm-03) picks it up automatically—but the raw observation also feeds into the general feed, where it’s visible to anyone asking “what’s happening near me?”
cm-03-trail-truth— Trail-specific observations route automatically to the conditions rillcm-09-citizen-sensor— Sensor data adds quantitative context to qualitative observationsvi-05-environmental-justice— Maps observation density against demographic vulnerability to reveal underreported areasag-01-event-discovery— Community-contributed event sightings feed the gathering calendar
Shared foraging intelligence with community-governed privacy. The group decides whether locations are shared at exact GPS, general area, or watershed level. Species observations include AI photo verification and phenological cross-reference—because the commons only works if the knowledge is trustworthy.
lat, lng (required); species, photo, quantity, conditions (optional); community membership token
Foraging map with privacy-graded precision, species confirmations, seasonal availability windows, contributor reputation, and AI verification status
Community contributions + AI species verification + phenological cross-reference from fl-06. Privacy governed by community rules.
Seasonal. Peak activity during fruiting seasons. Observations validated against phenological windows.
A foraging community in the Cascade foothills east of Seattle shares chanterelle observations at “fuzzy” resolution—1km radius, not exact coordinates. A contributor with high reputation posts a photo; the AI verification model confirms Cantharellus formosus. Phenology data confirms this matches the expected fruiting window for the region. The observation feeds into cm-10 (Harvest Reports) and back into fg-04 (Habitat Predictor), making the prediction model smarter without exposing anyone’s secret spot.
cm-10-harvest-reports— Closes the prediction feedback loop with ground-truth outcomesfl-06-phenology-tracker— Validates observation timing against accumulated degree-dayst-02-soil-profile— Soil type cross-reference strengthens habitat prediction accuracyfg-04-habitat-predictor— Community data trains the probability model for species occurrence
What the trail is actually like right now—not what the guidebook says. Muddy, washed out, downed trees, snow coverage, bridge damage, all reported by people who walked it today and cross-referenced with recent precipitation.
trail_id or lat, lng (required); condition_type, severity, photo (optional)
Trail condition summary: current reports, aggregate condition rating, recent weather context, contributor corroboration count, and report freshness
Community reports with provenance. Cross-referenced with NWS recent precipitation, USFS trail status where available. Free.
Real-time. Reports age out based on conditions—a “muddy” report after 3 dry days fades; a “bridge out” persists until resolved.
Before a Saturday hike on the Wildwood Trail in Forest Park, Portland, a runner queries cm-03. Three reports from the past 48 hours: “sections near Leif Erikson junction are ankle-deep mud,” “bridge at Balch Creek is passable but slippery,” and “upper sections are clear and dry.” Cross-referenced with 1.2 inches of rain two days ago, the mud reports have high plausibility. Stack with c-03 (trail network) for an alternate route that avoids the low-elevation creek crossings.
c-03-trail-network— Adds real conditions to the static trail map, enabling weather-aware route alternativesa-01-hyperlocal-weather— Forecast context helps interpret condition reports and predict improvementsa-08-movement-pattern— Matches trail difficulty to your current fitness and recovery statecm-01-community-observations— General observations about trail-adjacent areas (wildlife, hazards) enrich the picture
Community noise measurements using phone microphones, tagged with time, location, and source type. Builds a real noise map from actual decibel readings—not the modeled estimates from the Bureau of Transportation Statistics, but what your neighborhood actually sounds like at 2 AM on a Friday.
lat, lng (required); decibels (measured), source_type (traffic, construction, aircraft, music, etc.), duration (optional)
Noise map: heatmap of reported levels, time-of-day patterns, source type breakdown, statistical aggregates with outlier detection
Phone microphone readings + user source tags. Statistical outlier detection filters bad measurements. BTS noise contours as baseline comparison. Free.
Real-time accumulation. Aggregate patterns stabilize after ~50 reports per area. Individual readings decay in relevance over weeks.
A resident considering an apartment near SE Division Street in Portland queries cm-04. Community measurements show 68 dB average during weekday evenings (restaurant/bar noise) dropping to 42 dB after midnight, but spiking to 78 dB on Friday and Saturday nights until 1 AM. The BTS modeled estimate for this block is 55 dB—a number that averages away the temporal pattern. Stack with vl-06-comparable-sales and the noise pattern becomes a price signal: similar units on the quieter next block over rent for the same price, suggesting this one may be overvalued for its livability.
vl-06-comparable-sales— Noise levels modulate livability-adjusted property valuesvi-05-environmental-justice— Maps noise burden against demographic vulnerabilitysa-03-sleep-architecture— Correlates your sleep quality with ambient noise at your locationc-09-heatmap-generator— Renders community noise data as a spatial heatmap on the Cartograph stack
Community-reported water conditions—stream levels, clarity, algae blooms, flooding, dry streambeds. Especially valuable where USGS gauges are sparse, which is most places. The creek your kids play in probably doesn’t have a federal sensor on it.
lat, lng (required); water_body, condition (clarity, level, algae, odor), photo (optional)
Water condition map: recent reports per water body, condition trends, USGS cross-reference where overlap exists, confidence scores
Community reports. Cross-referenced with USGS NWIS stream gauges where overlap exists (free). Temporal plausibility check against recent precipitation.
Real-time. Water conditions change rapidly with weather. Reports more than 72 hours old are flagged as potentially stale.
Community members along Johnson Creek in SE Portland report green-brown water and a sulfurous smell after heavy rains. Three reports within 24 hours. USGS gauge data at the Milwaukie station confirms elevated flow. Cross-reference with vi-03 shows a brownfield site 800 meters upstream. Stack with vi-01 (water quality) for the institutional testing data and vi-05 (environmental justice) for the demographic context. This surfaces a signal that warrants investigation by the DEQ—community observations as the early warning system that triggers institutional response.
h-03-stream-gauge— USGS gauge data validates community water level reports where sensors existvi-01-water-quality— Institutional water testing adds regulatory context to community observationsvi-03-toxic-release— Proximity to contaminated sites explains anomalous water conditionsa-08-precipitation-patterns— Recent rainfall explains or predicts water condition changes
Where can I borrow a pressure canner? Who has extra tomato seedlings? Is there a repair cafe this Saturday? The local resource-sharing layer—tool libraries, seed exchanges, community gardens, food forests, skill shares—with community governance over listings and trust.
lat, lng (required); resource_type (tool, seed, skill, space), search_radius (optional)
Nearby resources: tool libraries, seed exchanges, community gardens, skill offers/requests, repair events, with contributor reputation and availability status
Community contributions with governance layer. Listings moderated by community rules. Free.
User-driven. Listings persist until marked fulfilled, expired, or removed by contributor. Seasonal patterns for seed and garden resources.
A new homeowner in Ballard, Seattle queries cm-06 looking for garden tools. Three results within a mile: a tool library at the Ballard P-Patch (available Tuesday and Saturday), a neighbor offering to lend a broadfork, and a community garden with shared compost bins. Stack with cu-01 (garden designer) to get a planting plan, then with cm-06 again to find the specific tools you’ll need and where to borrow them—a complete garden startup without buying anything.
cu-01-garden-designer— Planting plans generate tool needs that mutual aid can fulfillag-05-market-exchange— Farm stands and plant swaps are a commercial extension of mutual aidcm-08-neighborhood-pulse— Local bulletins surface new mutual aid opportunitiesag-04-work-party— Collective labor events draw from and feed the mutual aid network
The knowledge that experienced locals carry: where the spring dries up first, which hillside gets frost last, where deer cross the road, which fields flood in a ten-year storm. Oral tradition digitized—with attribution, provenance, and the respect the knowledge deserves.
lat, lng (required); topic (hydrology, wildlife, weather patterns, land history) (optional)
Local knowledge entries: long-form narratives with AI-extracted structured data, contributor attribution, community verification status, and historical cross-references
Community contributions (long-form narrative + voice transcription). AI structured extraction preserves the story while indexing the knowledge. Free.
Slow accumulation. Knowledge persists indefinitely. Verification through community corroboration over time.
A third-generation farmer on Sauvie Island near Portland contributes: “The field east of Reeder Road floods every time the Willamette crests 18 feet at the Portland gauge. Has since my grandfather’s time. FEMA doesn’t show it.” AI extraction tags this with a hydrological threshold (18 ft gauge reading), a spatial reference, and a temporal claim (multi-generational). Stack with h-03 (stream gauge) historical data to validate the threshold. Cross-reference with vi-05 (environmental justice) to see whether unmapped flood risk correlates with underserved communities. This is the data that doesn’t exist in any federal database—until someone writes it down.
h-03-stream-gauge— USGS gauge history validates community hydrological claimsfl-06-phenology-tracker— Phenological observations from locals ground-truth the computational modelt-05-land-cover— Historical land cover change adds quantitative backing to narrative land use storiesar-04-historical-timeline— Places local knowledge in the broader historical context of the area
Hyperlocal community sentiment and information—new business opening, construction starting, road closure, lost dog, community event. A structured, location-aware bulletin board that turns neighborhood gossip into searchable, tagged, geo-referenced intelligence.
lat, lng (required); radius (default 1 mile); event_type filter (optional)
Feed of neighborhood events: structured entries with type tags, location, recency, and contributor info. Filterable by category.
Community contributions with event-type tagging. Community governance for moderation. Free.
Real-time. Entries age naturally. Construction and road closure notices persist; lost-dog posts expire after resolution.
Someone considering renting in Seattle’s Fremont neighborhood queries cm-08 for the last 90 days. The pulse shows: a new café opening on 35th, construction starting on a 6-story mixed-use building on Stone Way, a community garden expanding its plots, and a recurring Sunday farmers market. Stack with cm-04 (noise witness) to check whether the construction generates complaints, and with vl-06 (comparable sales) to see if the new development is affecting nearby rents. The neighborhood pulse is the qualitative layer that numbers alone can’t capture.
ag-01-event-discovery— Structured events from the pulse feed into the formal events calendarvl-06-comparable-sales— Neighborhood changes correlate with property market movementscm-06-mutual-aid— New mutual aid resources surface through the pulse before formal listingcv-02-permit-tracker— Permit data adds institutional context to construction observations
The aggregation point for community-deployed environmental sensors—PurpleAir monitors, personal weather stations, soil moisture probes, sound level meters. The Internet of Community Things, with statistical validation and calibration tracking to keep the data honest.
lat, lng (required); sensor_type filter (air, weather, soil, noise); radius (optional)
Sensor data streams: real-time readings from nearby community sensors, calibration status, statistical validation scores, and comparison with institutional monitors
PurpleAir API (free), Weather Underground Personal Weather Stations (free), community-deployed hardware. Hardware provenance tracked. Statistical outlier detection.
Real-time streaming. Sensor readings typically update every 2–10 minutes depending on hardware.
A neighborhood association in SE Portland deploys three PurpleAir monitors after residents report persistent odors from an industrial area. cm-09 aggregates the readings and compares them to the nearest EPA AirNow station 2.3 miles away. The community sensors consistently show PM2.5 readings 40% higher than the official monitor. Statistical validation confirms the readings are consistent across all three sensors and the discrepancy is genuine, not sensor error. Stack with vi-05 (environmental justice screening) and the disparity between official monitoring and community measurement becomes a documented environmental equity case.
vi-04-air-quality— Institutional AQI data provides the baseline that community sensors can challenge or confirmvi-05-environmental-justice— Sensor disparities mapped against demographics reveal monitoring gapsa-01-hyperlocal-weather— Personal weather stations complement forecast models with ground truthc-09-heatmap-generator— Renders sensor networks as spatial heatmaps showing environmental gradients
Ground-truth outcomes from foraging expeditions—what people actually found, when, in what quantity, in what condition. Closes the feedback loop between habitat prediction models and reality. The prediction said chanterelles here; did anyone find them?
lat, lng (required); species, quantity, condition, photo (optional); linked prediction ID from fg-04 (optional)
Harvest log: species confirmed, quantity, location (privacy-gated), conditions, photo verification status, and prediction accuracy score if linked to a fg-04 prediction
Community contributions + AI photo verification + prediction model feedback. Cross-referenced with phenological data. Free.
Seasonal. Peak during fruiting seasons. Each report improves the prediction model for next season.
A forager near Gifford Pinchot National Forest logs a harvest report: “2.5 lbs chanterelles, excellent condition, old-growth Douglas fir stand at ~1,800 ft elevation.” Photo verified as Cantharellus formosus. The prediction model (fg-04) had assigned this area a 72% probability based on soil moisture, canopy density, and accumulated degree-days. The confirmed harvest bumps the model’s confidence for similar conditions. Over a season, dozens of harvest reports across the region tune the prediction model from theoretical to empirical—a collective intelligence that improves for everyone.
fg-04-habitat-predictor— Harvest outcomes feed directly back into prediction model trainingcm-02-foraging-commons— Reports enrich the shared foraging knowledge basefl-06-phenology-tracker— Actual harvest timing calibrates the phenological modelssa-04-nutrient-compass— Logged harvests can be nutritionally analyzed for dietary tracking
commune (together, in common)
How do people share what they’ve found? Commune is the collaboration and sharing layer—the tools for turning personal rill outputs into community resources. Share cards that make data beautiful, embeddable snippets for websites, persistent group channels with live data, collaborative dashboards, and the ability to gift a fully configured flow to someone as a link. This is where Slipstream stops being a solo tool and becomes a social one.
Shareable rill visualizations, embeddable data widgets, collaborative workspace state, multi-user dashboard configurations, gifting and onboarding flows, and community publication channels.
Internal: any rill output rendered as a share-ready artifact. Collaboration layer uses Supabase real-time for multi-user state. No external APIs—commune wraps Slipstream’s own output into shareable formats.
Without commune, every rill output is a private data point. With it, a soil profile becomes a shareable card on a community garden’s website. A foraging flow becomes a gift you send to a friend who just moved to the area. A collaborative dashboard becomes a neighborhood association’s living environmental monitor. co-05 (Gift a Flow) is especially powerful—the onboarding is the gift, meaning someone’s first experience with Slipstream is a fully configured, personally relevant dashboard that someone who cares about them assembled.
Stack co-04 (collaborative dashboard) + cm-09 (citizen sensors) + vi-04 (air quality) + a-01 (weather) + cm-04 (noise witness) for a neighborhood association in Seattle’s Georgetown district. Three board members co-edit the dashboard layout, pinning PurpleAir readings, official AQI, noise measurements, and weather. co-02 generates an embed snippet for the neighborhood website. co-06 publishes weekly summaries to the community feed. The dashboard is a living document of the neighborhood’s environmental conditions, maintained collectively—civic infrastructure built from rill compositions.
| ID | Name | Description |
|---|---|---|
| co-01 | Share Card | Beautiful share images from any rill output, optimized for social and messaging |
| co-02 | Embed Snippet | Iframe-embeddable rill visualizations for external websites and blogs |
| co-03 | Crew Channel | Persistent shared space for groups with live rill data embedded in conversation |
| co-04 | Collaborative Dashboard | Multi-user dashboard editing with role-based access and real-time sync |
| co-05 | Gift a Flow | Send a pre-configured dashboard as a link where the onboarding is the gift |
| co-06 | Community Broadcast | Publish observations and findings to your communis community feed |
Generates beautiful, shareable images from any rill output—soil profiles rendered as elegant cards, weather summaries as clean graphics, foraging finds as styled postcards. Designed for messaging apps and social feeds, not dashboards. The output is the artifact.
rill_output (any rill’s response data); template (card style); format (png, svg, webp)
Rendered image: styled card with data visualization, source attribution, Slipstream branding, and share-ready dimensions (1200×630 for social, 1080×1080 for square)
Internal. Wraps any rill output. Template library with M3 design tokens. Canvas/SVG rendering. No external API.
On demand. Generated at share time from current rill data.
A gardener in Portland’s Woodstock neighborhood queries t-02 (soil profile) for her backyard plot. The result shows well-drained Multnomah silt loam, pH 5.8, good organic matter. She generates a share card and texts it to her garden club: “Look what we’re working with this season.” The card renders as a beautiful graphic—soil type, drainage rating, pH gauge, organic matter percentage—with the location noted as “Woodstock, Portland” rather than exact coordinates. Data becomes conversation.
co-06-community-broadcast— Share cards become the visual format for community broadcastsco-05-gift-a-flow— The card is the preview of what the gift containsfm-04-dashboard-composer— Dashboard tiles can be exported as individual share cards
Turns any rill visualization into an iframe-embeddable widget for external websites and blogs. A neighborhood association’s site gets a live air quality gauge. A garden club’s blog gets an embedded phenology calendar. The data lives in Slipstream; the surface lives wherever it’s needed.
rill_id (required); config (dimensions, theme, refresh interval); permissions (public or token-gated)
Embed code (iframe HTML), hosted widget URL, optional API key for token-gated access, and preview render
Internal. Serves a lightweight version of any rill’s visualization. Hosted on Slipstream infrastructure.
Configurable. The embed polls the underlying rill at the specified interval (default: 15 minutes).
A community garden in Seattle’s Beacon Hill neighborhood wants to display current weather and soil conditions on their website. co-02 generates an embed snippet for a composite widget showing a-01 (weather) and t-02 (soil) data for the garden’s coordinates. The iframe is 400×300 pixels, updates every 30 minutes, and uses the garden’s green color theme. Visitors to the website see real-time growing conditions without leaving the page.
co-04-collaborative-dashboard— Dashboard layouts become embeddable compositescm-09-citizen-sensor— Community sensor networks get public-facing displaysfm-02-dataviz-engine— Shared visualization components render consistently across embed and app
A persistent shared space where groups can embed live rill data directly in conversation. Not a chat app—a data-native collaboration surface where the weather gauge, the soil profile, and the foraging map live alongside the discussion about what to plant this weekend.
channel_id (required); members (invite by link or username); pinned_rills (rill configurations to embed in the channel)
Shared channel view: threaded conversation with embedded rill cards, pinned data widgets, member roster, and activity timeline
Internal. Supabase real-time for multi-user state sync. Embedded rills update independently per their own cadence.
Real-time for conversation. Embedded rill data refreshes per each rill’s own cadence.
A beekeeping collective on Vashon Island creates a crew channel for their five apiaries. Pinned rills: a-01 (weather for each site), fl-06 (bloom status), and ap-02 (hive inspection window). Members post updates: “Checked Apiary 3 today, strong queen, good brood pattern.” The embedded bloom tracker shows Douglas fir coming into flow next week. The data and the conversation live together—context you’d otherwise have to assemble from five different apps.
co-04-collaborative-dashboard— Channels can link to shared dashboards for deeper data explorationcm-01-community-observations— Observations posted in channels can be promoted to the community feedco-06-community-broadcast— Channel highlights become community-wide broadcasts
Multi-user dashboard editing with role-based access and real-time sync. Multiple people can arrange rill tiles, configure thresholds, and annotate findings simultaneously. The bento box layout with collaborative editing—a shared operating picture for any group that needs one.
dashboard_id (required); collaborators (email or link); permissions (viewer, editor, admin)
Shared dashboard: bento grid of rill tiles, real-time cursor presence, edit history, comment threads on specific tiles, and version snapshots
Internal. Dashboard state in Supabase with real-time subscriptions. Embedded rills fetch their own data independently.
Real-time for layout changes and collaboration. Underlying rill data refreshes per each rill’s cadence.
A land trust in rural Clackamas County, Oregon creates a collaborative dashboard for their 2,000-acre conservation property. Three staff members have editor access. They arrange tiles: t-02 (soil), fl-02 (native plants), h-03 (stream gauge), c-02 (satellite overlay), and cm-07 (local ecological knowledge from neighboring farmers). The education director adds a comment on the stream gauge tile: “Flow dropped below 15 cfs—check if irrigation upstream increased.” The dashboard becomes the trust’s living operational view of their land.
co-02-embed-snippet— Collaborative dashboards can be embedded on external sitesco-05-gift-a-flow— A dashboard can be gifted to introduce someone to a complete data viewfm-04-dashboard-composer— Personal dashboards can be promoted to collaborative onesco-03-crew-channel— Dashboards link to channels for context-rich discussion
Send a pre-configured dashboard as a link. The recipient clicks, enters their location, and immediately has a working, personally relevant rill composition—no setup, no learning curve. The onboarding is the gift. Someone’s first experience with Slipstream is something a person who cares about them built for them.
flow_config (rill composition + layout); personalization_prompts (what to ask the recipient: location, interests, birth data for astrology); message (optional personal note)
Share link that opens a guided onboarding: personal message, then 2–3 setup questions, then a fully rendered dashboard customized to the recipient’s answers
Internal. The gift is a template; the data is fetched fresh for the recipient’s location. No cost beyond the underlying rills.
On open. The gift link is persistent; data renders fresh each time the recipient visits.
For his mom’s birthday, Devin configures a flow: as-03 (moon tracker) + as-07 (horoscope narrator) + sp-02 (Warriors game alerts) + a-01 (weather). He adds a personal note: “Happy birthday, Mom. Open this every morning.” She clicks the link, enters her birth date/time and location, and immediately has a morning dashboard: today’s moon sign, a personalized transit reading, the next Warriors game, and the weather. She didn’t install anything. She didn’t configure anything. She just received a gift that keeps refreshing.
in-05-gift-receiver— The optimized flow for opening and onboarding into a gifted dashboardco-04-collaborative-dashboard— Gifted flows can be collaborative if the giver enables itas-07-horoscope-narrator— The daily narrative makes the gift feel personal every single morningsp-02-sports-pulse— Sports alerts add the specific, personal touch that makes it “her” dashboard
Publish observations, findings, and composed insights to your communis community feed. Not a social post—a structured data publication with provenance, source attribution, and the ability for others to fork and explore the underlying rill stack.
content (rill output, share card, or composed narrative); community_id (which community to publish to); privacy_level (public, community-only)
Published entry in community feed: attributed, timestamped, with linked rill sources that viewers can explore
Internal. Wraps any rill output or composition. Community governance determines moderation and visibility.
On publish. The broadcast is a snapshot; linked rills update independently for anyone who explores further.
A birdwatcher in the San Juan Islands composes fa-01 (bird migration radar) + a-01 (weather) + cm-01 (community observations) and discovers a pattern: a wave of western tanagers arrived three days earlier than historical average, correlating with unseasonably warm winds from the south. She publishes this finding via co-06 to her birding community. Other members can see the broadcast, click through to explore the underlying rill data, and add their own observations. The finding becomes community knowledge, not a private insight.
cm-01-community-observations— Observations in broadcasts feed back into the community data layerco-01-share-card— Broadcasts use share cards as their visual formatco-03-crew-channel— Channel discussions can be broadcast to the wider community
agora (gathering place, marketplace)
Where do people gather, and what brings them together? Agora covers community-organized events tied to place and season—farmers markets, work parties, skill shares, BioBlitz surveys, seasonal celebrations, and the informal gathering rhythms that define a neighborhood’s social fabric. These aren’t entertainment events (that’s spectaculum); these are the gatherings where neighbors become community.
Local events from multiple sources, seasonal community calendars tied to natural cycles, skill-sharing exchanges, collective labor events (work parties, barn raisings), farm stands and plant swaps, and community biodiversity survey events.
Eventbrite API (free tier), community contributions, iNaturalist API (BioBlitz events, free), local venue calendar feeds, NWS forecast data for weather-optimized scheduling. Most agora rills are community-sourced rather than API-driven.
A farmers market is an event. Stack it with fl-06 (phenology) and you know what’s in season. Add a-01 (weather) and it becomes a weather-informed outing decision. Layer ag-04 (work party) with a-01 and t-02 (soil) and you schedule the garden cleanup for the Saturday when the soil is dry enough to turn. The most powerful composition is ag-06 (BioBlitz coordinator) stacked with the full biosphere domain—turning a community survey event into a citizen science expedition with institutional-quality data collection.
Stack ag-04 + a-01 + t-02 + cm-06 for a community garden work party in Seattle’s Rainier Beach. The organizer sets the task (turn beds, plant starts) and a two-week window. a-01 forecasts a dry stretch starting Thursday. t-02 confirms the soil type drains well after 48 hours without rain. The system suggests Saturday morning as optimal—two dry days will make the soil workable. cm-06 checks tool availability at the nearby lending library. The invitation goes out with weather context: “Soil will be ready—bring gloves, shovels available at the tool library.”
| ID | Name | Description |
|---|---|---|
| ag-01 | Event Discovery | Local community events from multiple sources, filterable by interest and distance |
| ag-02 | Seasonal Calendar | Community gathering calendar tied to natural cycles and phenological triggers |
| ag-03 | Skill Share | Teaching and learning exchange network tied to seasonal availability |
| ag-04 | Work Party Organizer | Collective labor events with weather-optimized scheduling and tool coordination |
| ag-05 | Market & Exchange | Farm stands, pop-ups, plant swaps, and tool shares in your area |
| ag-06 | BioBlitz Coordinator | Community biodiversity survey events with iNaturalist bridge for data collection |
Aggregates community events from Eventbrite, local venue calendars, and community contributions into a single feed filterable by interest, distance, and date. Not the big-ticket concerts—the neighborhood potlucks, volunteer cleanups, and seed swaps that don’t make it onto mainstream platforms.
lat, lng (required); radius, category (gardening, cleanup, social, educational), date_range (optional)
Event list: title, location, time, organizer, category, source (Eventbrite vs. community-contributed), and weather forecast for outdoor events
Eventbrite API (free tier), community contributions via cm-08 (neighborhood pulse), local venue iCal feeds. Free.
Daily. Eventbrite polled nightly; community contributions appear in real-time.
A new resident in Tacoma’s Hilltop neighborhood queries ag-01 filtered by “gardening” and “food” within 3 miles. Results: a community garden open house this Saturday, a seed swap at the library next week, and a canning workshop at the neighborhood center. None of these were on Eventbrite—they came through community contributions. Stack with a-01 (weather) and the Saturday garden event shows “partly cloudy, 58°F, no rain expected”—perfect garden weather.
a-01-hyperlocal-weather— Weather context for outdoor events transforms discovery into decisioncm-08-neighborhood-pulse— Pulse entries tagged as events auto-feed the discovery calendarsp-01-live-music— Entertainment events complement community events for a complete local calendarag-02-seasonal-calendar— Discovery events slot into the broader seasonal rhythm view
A community gathering calendar tied to natural cycles rather than arbitrary dates. Planting parties when the soil warms. Harvest festivals when the crops come in. Mushroom walks when the rains return. The calendar follows the land, not the Gregorian grid.
lat, lng (required); interests (gardening, foraging, astronomy, birding) (optional)
Seasonal timeline: upcoming community events mapped to phenological triggers, solstice/equinox markers, growing season milestones, and recurring community rhythms
Composed from fl-06 (phenology), t-01 (helio), seasonal community event patterns. Phenological triggers computed from degree-days.
Weekly. Phenological triggers update as degree-days accumulate. Community events added in real-time.
A community garden coordinator in Eugene, Oregon opens ag-02 in early March. The seasonal calendar shows: “Soil temperature approaching 50°F—pea planting window opens in ~10 days,” “Spring equinox potluck (community tradition) March 20,” and “First bloom forecast: ornamental cherry, ~March 15.” She schedules the garden’s spring kickoff to coincide with the soil temperature threshold, not an arbitrary weekend. The calendar becomes a shared reference for the garden’s rhythm.
fl-06-phenology-tracker— Phenological triggers drive the seasonal calendar’s event suggestionst-01-helio-study— Day length and solar angle mark the astronomical milestonesag-04-work-party— Seasonal work events slot into the calendar at their optimal timingas-06-seasonal-astrology— Astrological season markers overlay the natural cycle calendar
A teaching-and-learning exchange network tied to seasonal availability. Soap making in winter when the herbs are dried. Grafting workshops in early spring when the sap rises. Canning classes during harvest season. The skills follow the seasons, and the teachers are your neighbors.
lat, lng (required); skill_category (gardening, preservation, craft, repair); role (teach or learn) (optional)
Skill exchange listings: offers and requests, seasonal relevance, instructor reputation, prerequisites, and upcoming session schedule
Community contributions. Seasonal relevance computed from phenology and calendar. Instructor reputation from cm-01 trust layer. Free.
User-driven. Listings appear as contributors post them. Seasonal suggestions refresh weekly.
In late September on Whidbey Island, a homesteader queries ag-03 for “preservation” skills. Three results: a neighbor offering a pressure canning workshop (she’s canned for 30 years, high reputation), a community college running a fermentation class, and someone requesting help learning to smoke salmon. The seasonal timing is perfect—salmon runs are peaking, the apple harvest is in, and root vegetables are ready to store. Stack with cm-06 (mutual aid) to find the shared tools needed for each skill.
cm-06-mutual-aid— Skills need tools, and mutual aid provides themag-02-seasonal-calendar— Skill offerings slot into the seasonal rhythm naturallyag-04-work-party— Work parties are skill shares in action—learning by doing together
Collective labor events with weather-optimized scheduling, tool coordination from the mutual aid network, and task breakdown for groups. The modern barn raising—a community garden needs beds turned, a trail needs clearing, a neighbor needs help with firewood.
task (description), location (lat/lng), date_window (flexible range), tools_needed, crew_size (optional)
Event plan: optimal date (weather-informed), tool availability check, task breakdown, sign-up sheet, and weather briefing for the selected date
Composed: a-01 (weather forecast), cm-06 (tool availability), community sign-ups. No external API beyond weather.
Event-driven. Weather forecast updates daily as the work party date approaches.
A neighborhood association in Portland’s Alberta Arts District organizes a spring cleanup for a neglected intersection planting strip. The organizer enters the task (“clear invasive blackberry, plant native shrubs”), location, and a two-week window. ag-04 checks the forecast and suggests the second Saturday—a dry day after two days without rain, so the soil won’t be too muddy for digging. Tool check via cm-06: loppers and shovels available at the nearby tool library, but they’ll need someone to bring a wheelbarrow. Six neighbors sign up. The event page shows: task breakdown, weather forecast, tool checklist, and a map.
a-01-hyperlocal-weather— Optimal scheduling based on precipitation, temperature, and windcm-06-mutual-aid— Tool availability check integrated into event planningt-02-soil-profile— Soil type determines how many dry days are needed before working the groundag-02-seasonal-calendar— Work parties slot into the seasonal calendar for community visibility
Farm stands, pop-up markets, plant swaps, tool shares, and barter events in your area. The informal economy of a community—the places where surplus zucchini, extra seedlings, and homemade jam change hands, often at a folding table in someone’s driveway.
lat, lng (required); type (farm stand, swap, market, barter); radius (optional)
Exchange listings: location, type, schedule (recurring or one-time), what’s available (if listed), and community ratings
Community contributions. Seasonal availability context from fl-06 (phenology). USDA Farmers Market Directory (free) for established markets.
Daily. Recurring markets maintain persistent listings. Pop-ups and one-time events added by contributors.
A family in Olympia, Washington queries ag-05 on a Saturday morning. Results: the downtown farmers market (weekly, 10 AM–3 PM), a plant swap at a neighbor’s yard (one-time, this afternoon), and a farm stand on the road to Tumwater with fresh eggs and early greens. Stack with fl-06 and the market listing shows seasonal context: “First asparagus of the season typically available this week at South Sound markets.” The exchange layer turns seasonal knowledge into a shopping list.
fl-06-phenology-tracker— Seasonal availability context for what’s likely at markets this weekcm-06-mutual-aid— Markets and exchanges extend the mutual aid network into commercesa-04-nutrient-compass— Nutritional context for seasonal market finds
Organizes community biodiversity survey events with an iNaturalist bridge for institutional-quality data collection. A Saturday morning becomes a citizen science expedition—structured observation protocols, species identification support, and data that flows back to the global biodiversity record.
location (survey area polygon); date; focus (all taxa, birds, plants, fungi, invertebrates); participants (optional)
BioBlitz plan: survey area map, species checklist for expected taxa, observation protocols, iNaturalist project link, participant assignments, and post-event species tally
iNaturalist API v1 (free) for project creation and observation sync. GBIF for expected species lists. Community sign-ups. fa-01 for migration timing.
Event-driven. Species lists update as observations are submitted during the BioBlitz.
A neighborhood nature group in Bainbridge Island organizes a spring BioBlitz for a 50-acre wetland preserve. ag-06 generates a survey plan: expected species based on GBIF records and current phenology (wildflowers in bloom, early migrants arriving per fa-01), suggested transect routes, and observation protocols for beginners. An iNaturalist project is auto-created. Twelve participants spend Saturday morning documenting everything they find. By afternoon, 340 observations are submitted, 89 species identified, including a sensitive orchid species not previously recorded at this site. The data flows to iNaturalist, GBIF, and back into Slipstream’s biodiversity layer.
fa-01-bird-migration— Migration timing suggests optimal BioBlitz dates for maximum species diversityfl-02-native-plant-atlas— Expected native species checklist for the survey areacm-01-community-observations— BioBlitz observations feed the community intelligence layerco-06-community-broadcast— Post-event species tally broadcast to the wider community
spectaculum (spectacle, public show)
What’s happening tonight? Spectaculum covers the cultural calendar—concerts, sports, gallery openings, film screenings, food pop-ups, and theater. These are the events produced by the culture industry, not by your neighbors (that’s agora). This is the family that makes Slipstream relevant to people who will never forage a mushroom but absolutely want to know about the jazz show at the corner bar or when the Warriors play next.
Live music and concert listings, professional sports schedules and scores, gallery and museum exhibitions, independent film screenings and festivals, food and drink events, and performing arts (theater, dance, comedy, spoken word).
Songkick API (free tier), Bandsintown API (free), TheSportsDB (free, no key), ESPN hidden API (undocumented but widely used), Eventbrite API (free tier), Artsy API (exhibitions), local venue calendar feeds, community contributions.
Entertainment alone is a calendar. Stack sp-02 (Sports Pulse) with as-03 (Moon Tracker) and as-07 (Horoscope Narrator) and you have a morning briefing that covers the Warriors game tonight, the moon sign, and a personalized daily reading—a product someone opens every day. Layer sp-01 (Live Music) with a-01 (weather) and outdoor venue listings gain weather context. The spectaculum family is what makes Slipstream a daily companion, not just a research tool.
Stack sp-02 + as-03 + as-07 + a-01 for a personalized morning briefing. The Warriors play the Lakers tonight at 7:30 PM on TNT. The moon is in Libra—harmony, balance, relationships. Jupiter trines her natal Moon today at 1.2° orb: “An expansive, generous emotional energy. Good day for connecting with family.” Weather: 62°F, partly cloudy, golden hour at 5:23 PM looks exceptional. That’s not a dashboard. That’s something someone opens every morning because it’s hers.
| ID | Name | Description |
|---|---|---|
| sp-01 | Live Music Radar | Concerts, shows, DJ sets, and open mics nearby, filterable by genre and venue |
| sp-02 | Sports Pulse | Game schedules, scores, and alerts for teams you follow |
| sp-03 | Gallery & Exhibition | Current art shows, museum exhibitions, opening receptions, and closing-soon alerts |
| sp-04 | Film & Screening | Independent cinema, outdoor screenings, film festivals, and repertory programming |
| sp-05 | Food & Drink | Pop-up dinners, wine tastings, brewery releases, food truck rallies, and restaurant week |
| sp-06 | Performing Arts | Theater, dance, comedy, spoken word, poetry slams, and storytelling nights |
Concerts, shows, DJ sets, and open mics near you, filterable by genre, venue, price range, and distance. Not Ticketmaster’s algorithm—a clean feed of what’s actually happening, including the tiny venue shows that only exist on a flyer taped to a telephone pole.
lat, lng (required); genre, radius, price_range, date_range (optional)
Show listings: artist, venue, date/time, ticket price, genre tags, and source (API vs. community-contributed)
Songkick API (free tier), Bandsintown API (free), Eventbrite music events, community contributions. Free.
Daily. API sources polled nightly; community contributions appear in real-time.
A couple in Capitol Hill, Seattle wants to go out on Friday. sp-01 filtered by “jazz” and “under $25” within 3 miles: a trio at the Royal Room (free), a vocalist at Tula’s ($20 cover), and a community-contributed listing for a jam session at a house venue in the Central District. The Songkick listings cover the established venues; the community contribution surfaces the one they’d never find otherwise. Stack with a-01 (weather) for walking-weather context if they’re deciding between outdoor and indoor options.
a-01-hyperlocal-weather— Weather context for outdoor venues and the walk to the showag-01-event-discovery— Music events complement community events for a complete local calendarki-03-route-planner— Transit and walking routes to the venuesp-05-food-drink— Dinner before the show, dessert after—the evening composed
Game schedules, scores, and alerts for teams you follow. Not a sports app—a rill. Subscribe to the Warriors, get a pre-game alert three hours before tipoff, skip the noise of live-score buzzing, and get the final score with top scorer and next game. Clean, minimal, respectful of your attention.
team_subscriptions (team name, league); alert_config (pre-game timing, live score on/off, final score on/off)
Per-team: next game card (opponent, time, venue, broadcast), season record, recent results, upcoming schedule, and triggered alerts per subscription config
TheSportsDB (free, no key). ESPN hidden API (undocumented, widely used). Ball Don’t Lie API (free NBA data). Free.
Schedule: daily. Live scores: on request (not push by default—user controls alert frequency).
Devin’s mom subscribes to the Golden State Warriors. Pre-game alert at 3 hours: “Warriors vs Lakers tonight, 7:30 PM, Chase Center, TNT. GSW 34–18, 3rd in West, four-game win streak.” She doesn’t want live-score buzzing during the game. After it ends: “Warriors 112, Nuggets 105. Curry: 34 pts, 8 ast. Next: vs Celtics, Friday 5:30 PM.” That’s the entire sports experience she wants. No ESPN app with 47 notification categories. Just her team, her way.
as-07-horoscope-narrator— Game day meets daily transit reading in the morning digestas-03-moon-tracker— Moon sign and sports alerts combine into a complete personal briefingco-05-gift-a-flow— A sports + astrology morning dashboard is a birthday gift that keeps givinga-01-hyperlocal-weather— Weather matters for outdoor sports and tailgating
Current art shows, museum exhibitions, opening receptions, closing-soon alerts, artist talks, and free admission days. The cultural calendar for the art-curious—because the best show in the city is always the one you almost missed because nobody told you it was closing next week.
lat, lng (required); radius, medium (painting, sculpture, photography, mixed), type (opening, closing-soon, free-day) (optional)
Exhibition listings: venue, show title, artists, dates, admission, special events (openings, artist talks), and closing countdown
Artsy API (exhibitions, free tier), local museum iCal feeds, community contributions, Eventbrite cultural events. Free.
Weekly. Exhibition data is relatively stable; closing-soon alerts trigger at 2-week and 1-week thresholds.
An art-lover in Portland queries sp-03 filtered by “closing-soon.” Results: a photography show at Blue Sky Gallery closing in 5 days, the Portland Art Museum’s contemporary wing has free admission this Thursday, and a community-contributed listing for a studio open house in the Pearl District this weekend. The closing-soon alert is the real value—static exhibition dates buried on museum websites become actionable urgency signals.
sp-01-live-music— Gallery openings with live music combine into a cultural eveningsp-05-food-drink— Dinner near the gallery completes the outingki-03-route-planner— Walking route between gallery and dinner that passes through interesting streets
What’s playing at the indie theaters, film festivals, outdoor screenings, and repertory houses. Not the AMC showtimes—the stuff you’d miss if nobody told you: the 35mm print of Stalker at the Clinton Street Theater, the outdoor screening in the park, the short film festival at the community center.
lat, lng (required); type (repertory, outdoor, festival, documentary); radius (optional)
Screening listings: film, venue, format (35mm, digital, outdoor), date/time, admission, and festival context if applicable
Local theater calendar feeds, Letterboxd lists, film festival APIs, community contributions. Outdoor screenings checked against a-01 weather.
Daily. Theater schedules change weekly; festival programs update during festival season.
A film nerd in Portland queries sp-04 for the weekend. Results: a 35mm screening of Blue Velvet at the Hollywood Theatre (Saturday 9 PM), an outdoor documentary screening at Cathedral Park (Sunday 8 PM, weather-permitting—a-01 says clear skies, 65°F), and the Portland International Film Festival’s shorts program at Cinema 21. The outdoor screening gets a weather confidence badge: “Clear skies forecast, bring a blanket.”
a-01-hyperlocal-weather— Weather viability for outdoor screenings, with rain-risk warningssp-05-food-drink— Dinner and a movie, composed from local datat-01-helio-study— Sunset time determines when outdoor screenings can begin
Pop-up dinners, wine tastings, brewery releases, food truck rallies, restaurant week, and chef collaborations. The ephemeral food calendar—events that exist for one night, one weekend, or one season, and vanish before the food bloggers write about them.
lat, lng (required); type (pop-up, tasting, release, festival, truck); cuisine, radius (optional)
Food event listings: name, location, date/time, type, cuisine, price range, and whether reservations are required
Community contributions, Eventbrite food/drink events, restaurant social media feeds. Free.
Daily. Ephemeral events by nature—freshness matters most in this family.
A foodie in San Francisco queries sp-05 for the weekend. Results: a sake tasting at a new izakaya in Japantown (Saturday, $45), a food truck rally at Fort Mason (Sunday, free entry), and a community-contributed listing for a Burmese pop-up dinner in a Sunset District garage (Saturday, $30, 20 seats, book now). The pop-up listing is the gem—it exists nowhere else, posted by the chef two hours ago. Stack with sp-01 (live music) and the izakaya listing notes a jazz trio playing during the tasting. The evening composes itself.
sp-01-live-music— Dinner + music pairings compose naturallyag-05-market-exchange— Farmers markets are where food events and community exchange overlapsa-04-nutrient-compass— Nutritional context for adventurous eating (if you want it)
Theater, dance, comedy, spoken word, poetry slams, storytelling nights, and community performances. Live art that isn’t music—the fringe festival show, the improv night, the one-woman play at the black box theater around the corner from your apartment.
lat, lng (required); type (theater, dance, comedy, spoken-word); radius (optional)
Performance listings: show title, company/artist, venue, dates, ticket price, and run length (closing-soon flag for limited runs)
Local venue calendar feeds, community contributions, Eventbrite performing arts events. Free.
Weekly. Theater runs are multi-week; comedy and spoken word are weekly recurring. Closing-soon alerts trigger at 1-week threshold.
A Seattle resident queries sp-06 for comedy. Results: an improv night at Unexpected Productions (weekly, $15), a stand-up showcase at the Comedy Underground (Saturday, $20), and a community-contributed listing for a storytelling night at a Beacon Hill bar (free, monthly). The storytelling night doesn’t appear on any ticketing platform—it was posted by the organizer this morning. Stack with sp-05 (food and drink) for dinner options near the venue.
sp-05-food-drink— Dinner and a show, the classic compositionsp-01-live-music— A complete evening: dinner, show, music afterag-01-event-discovery— Community performances overlap with grassroots gathering events
salus (health, safety, well-being)
Every other family processes data about the world. Salus processes data about you—your circadian rhythm, your sleep architecture, your nutrition, your breath, your movement, your body’s readiness for the day. This is the most personal family in Slipstream, and it carries the strongest privacy guarantees: local-first processing, per-rill consent, full data export and deletion, and a clear line between health information and medical advice. Slipstream is not a doctor. It surfaces patterns; you interpret them.
Circadian rhythm and chronotype, wearable biometrics (HR, HRV, SpO², skin temperature), sleep staging and debt, nutritional intake and gaps, breathwork protocols with biofeedback, ultradian and infradian rhythms, botanical wellness mapping, movement patterns and diversity, composite readiness scoring, and structured health journaling.
Solar ephemeris (pvlib/astral, free computation), Open Wearables or Terra API (wearable data normalization, free/open-source), Apple HealthKit, Google Health Connect, Garmin/Oura/Whoop APIs (device-linked, free), USDA FoodData Central (380K+ foods, free, CC0), Open Food Facts (3M+ international products, free), NIH Office of Dietary Supplements (free). Most salus rills run on pure computation + user-owned device data.
Salus is the personal lens through which other families become actionable. Stack sa-01 (Circadian Compass) with t-01 (Helio Study) and your body clock overlays on the sun’s actual path at your latitude. Add sa-03 (Sleep Architecture) to vi-04 (Air Quality) and you correlate respiratory symptoms with AQI spikes. Layer sa-08 (Movement Pattern) with c-03 (Trail Network) and your exercise history maps onto the trail system: “You’ve explored 12% of trails within 5 miles. Here are three that match your current fitness.” The compound rill sa-09 (Recovery Score) sees sleep AND nutrition AND cycle phase AND barometric pressure AND training load—the intelligence that no single wearable can provide.
Stack sa-09 + sa-01 + sa-03 + sa-06 + a-01 for a personal morning briefing. “You slept 6.8 hours with good deep sleep but short REM. Recovery score: 72. Your ultradian peak hits at 9:45 AM—schedule your hardest work then. Pressure dropping—if you’re migraine-sensitive, stay hydrated. Sunrise was at 6:52; your light exposure window opens now.” This is the personal operating system: your body’s data, composed with the environment’s data, surfacing signals that help you make better decisions about your day.
| ID | Name | Description |
|---|---|---|
| sa-01 | Circadian Compass | Personal circadian rhythm from latitude, chronotype, and season |
| sa-02 | Pulse Reader | Wearable data gateway normalizing HR, HRV, SpO², and activity across devices |
| sa-03 | Sleep Architecture | Sleep stages, efficiency, debt accumulation, and circadian alignment |
| sa-04 | Nutrient Compass | Nutritional intelligence from USDA FoodData Central with gap analysis |
| sa-05 | Breath Pacer | Guided breathwork with real-time biofeedback and resonant frequency estimation |
| sa-06 | Body Clock | Ultradian (90-minute) and infradian (menstrual, seasonal) rhythm tracking |
| sa-07 | Botanical Wellness Bridge | Evidence-graded botanical compound mapping to physiological states |
| sa-08 | Movement Pattern | Activity classification, movement diversity scoring, and VO² max trends |
| sa-09 | Recovery Score | Compound readiness score (0–100) from HRV, sleep, nutrition, and cycle phase |
| sa-10 | Health Journal | Structured health logging with auto-attached physiology and pattern detection |
Computes your ideal circadian rhythm based on latitude, longitude, season, and chronotype. Not generic—calibrated to where you actually are and when the sun actually rises. Your melatonin window shifts with the seasons the way your body wants it to; most apps never tell you this.
lat, lng (required); chronotype (from MCTQ questionnaire or inferred from 2+ weeks of sleep data) (optional)
24-hour body clock: optimal melatonin window, cortisol peak prediction, ideal sleep/wake times, light exposure windows, social jetlag estimate, and seasonal shift animation
Solar ephemeris (pvlib/astral + latitude). Pure computation, no API. Free. Chronotype from MCTQ or Munich Chronotype Questionnaire (public domain).
Daily. Circadian timing shifts with day length—recalculated each day based on solar position at your location.
A night owl in Seattle opens sa-01 in mid-January. At 47.6°N latitude, sunrise is 7:52 AM, and her computed melatonin window opens at 9:15 PM—45 minutes earlier than she usually goes to bed. The social jetlag estimate shows she’s 38 minutes misaligned from her solar noon. The visualization overlays her cortisol peak (predicted at 8:30 AM) on the sun’s path, showing she should get light exposure within 30 minutes of waking. Stack with t-01 (Helio Study) and the body clock maps onto the solar geometry—a richer picture than any generic sleep app provides.
t-01-helio-study— Solar position provides the physical basis for circadian computationsa-03-sleep-architecture— Actual sleep timing vs. circadian ideal reveals alignment or driftsa-06-body-clock— Ultradian rhythms nest inside the circadian frameworkas-03-moon-tracker— Lunar cycle overlaid on circadian rhythm for those who track both
The bridge between your wearable devices and Slipstream. Ingests heart rate, HRV, SpO², skin temperature, and activity data from any connected device and normalizes it into a device-agnostic schema. The data gateway for the entire salus family—local-first by design.
Device connection via OAuth per platform. Supported: Apple HealthKit, Google Health Connect, Garmin, Oura, Whoop, Fitbit
Normalized time-series: HR, HRV (RMSSD), SpO², skin temperature, steps, active minutes, respiratory rate. Device-agnostic format.
Open Wearables (open-source, free, preferred) or Terra API (free tier). Individual vendor APIs free with device ownership. OAuth per platform.
Continuous sync when connected. Batch sync on device unlock. Data processed locally by default.
A runner in Portland connects her Garmin watch and her partner’s Oura ring. sa-02 normalizes both data streams into the same schema—despite coming from different hardware, different sampling rates, and different sensor configurations. Her Garmin provides higher-resolution activity data; the Oura provides more detailed sleep staging. The normalized output feeds sa-03, sa-08, and sa-09 without either downstream rill needing to know which device generated the data. Privacy architecture: raw biometric data stays on device; only derived metrics leave unless cloud sync is explicitly enabled.
sa-03-sleep-architecture— Sleep staging requires the normalized HRV and activity streamsa-08-movement-pattern— Activity classification uses the normalized accelerometer and HR datasa-09-recovery-score— The compound score depends on Pulse Reader as its data foundationsa-05-breath-pacer— Real-time HRV during breathwork requires the live data stream
Goes beyond “you slept 7 hours” to reveal sleep stage distribution, consistency, debt accumulation, and alignment with your circadian phase. Was your deep sleep low because you ate late, because the pressure dropped, or because your HRV was already suppressed? The compound connections make sleep legible.
Wearable sleep data via sa-02; circadian phase from sa-01; user-reported sleep logs (optional)
Sleep stage breakdown (REM/deep/light/awake), sleep efficiency %, sleep debt accumulator, consistency score, circadian alignment gauge, 90-minute cycle completion count
Compound rill: sa-02 raw wearable data + sa-01 circadian phase + user logs. Pure computation on ingested data. Free.
Daily. Recalculated each morning from overnight data. Sleep debt accumulator updates rolling.
A consultant in San Francisco checks sa-03 on a Monday morning. Sleep efficiency: 81% (below her 87% average). Deep sleep: 42 minutes (usually 65). REM: normal. The circadian alignment score shows she went to bed 90 minutes past her melatonin window. Sleep debt accumulator: 4.3 hours over the last week. Stack with a-01 (weather) and the barometric pressure dropped 8 mb overnight—correlated with her history of lighter sleep during pressure changes. The pattern is visible only because multiple data streams converge.
sa-01-circadian-compass— Circadian phase determines whether sleep timing was aligned or drifteda-01-hyperlocal-weather— Barometric pressure correlates with sleep quality for sensitive individualssa-09-recovery-score— Sleep quality is a primary input to daily readinesscm-04-noise-witness— Ambient noise at your location may explain poor sleep efficiency
Full nutritional intelligence from the world’s largest public food composition database. What you ate, what it contains, what you’re missing. Turns 380,000 USDA food records into personal nutritional awareness—including a forage-aware mode that knows those chanterelles you found have 3× the vitamin D of store-bought mushrooms.
food_items (text search or barcode scan); quantity; optional: linked forage journal entries from fg-06
Per-food nutrient breakdown (macro + micro), daily intake tracking, nutrient gap heatmap, DRI percentage coverage, and wild vs. store-bought comparison when applicable
USDA FoodData Central API (380K+ foods, free, CC0, 1000 req/hr, instant key). Open Food Facts (3M+ international products, free, barcode scan). Free.
On query. USDA database updates quarterly. Daily tracking accumulates per-user.
A forager near the Oregon coast logs a harvest of chanterelles and huckleberries. sa-04 cross-referenced with fg-07 (foraging nutrition) shows the chanterelles provide 40% of daily selenium and the huckleberries are dense in anthocyanins. The nutrient gap heatmap shows she’s been low on magnesium for the past week. The wild food data from USDA FoodData Central is more specific than generic “mushroom” entries—it knows Cantharellus formosus specifically. Nutrition becomes place-based knowledge.
fg-07-foraging-nutrition— Wild food nutritional profiles compared to store-bought equivalentssa-07-botanical-wellness— Nutritional gaps may suggest botanical supplementation with evidence gradingsa-09-recovery-score— Nutrition status feeds the daily readiness compoundag-05-market-exchange— Seasonal market finds become nutritionally contextualized
Guided breathwork with real-time biofeedback—not just a timer but a responsive system that adapts breathing patterns to your current physiological state. HRV responds to controlled breathing within 60 seconds. This rill closes the loop: reads your state, guides you through a protocol, and shows you the effect in real time.
protocol (box breathing, 4-7-8, coherence, Wim Hof); optional: live HR/HRV from sa-02; optional: CO² tolerance test results
Guided breath animation, real-time HRV trace during session, pre/post HRV comparison, resonant frequency estimation, coherence score, session history
Pure computation + curated protocol library. Optional real-time HRV from sa-02. Works without a wearable (timer-only mode). Free.
Real-time during session. Session records stored for trend analysis. Resonant frequency refines with each session.
Before a presentation, a product manager in Seattle opens sa-05. Her current HRV (via Oura ring, read through sa-02) is 38 ms—below her 52 ms average, indicating stress. The rill suggests coherence breathing at her estimated resonant frequency of 5.5 breaths per minute. After 8 minutes, HRV has climbed to 48 ms. The post-session card shows the real-time trace—the HRV amplitude increasing as her breathing settled into the resonant frequency. That’s biofeedback without clinical equipment.
sa-02-pulse-reader— Live HRV data enables real-time biofeedback during sessionssa-09-recovery-score— Breathwork sessions can improve recovery score metricssa-01-circadian-compass— Evening breathwork timed to the melatonin window enhances sleep onset
The rhythms inside the circadian cycle and beyond it—90-minute ultradian alertness waves, menstrual cycle phases, and seasonal energy patterns. Most people know about circadian rhythms but almost nobody tracks their ultradian peaks, the ~90-minute BRAC cycle that governs when you’re sharp and when you should rest.
HRV/activity patterns from sa-02; sleep cycle data from sa-03; optional: menstrual cycle tracking (user input); seasonal model computed from historical patterns
Ultradian cycle map (90-minute peaks/troughs), menstrual cycle phase with recommendations, seasonal energy pattern (SAD risk), peak performance windows, and “when to do what” scheduler
Compound: sa-02 + sa-03 + user input. Pure computation on ingested data. Needs ~2 weeks of data to model ultradian rhythms. Free.
Daily recalculation. Ultradian model refines as data accumulates. Seasonal pattern emerges over months.
A freelance designer in Portland has been using Slipstream for three weeks. sa-06 has now identified her ultradian pattern: peak focus at 9:30 AM, a trough at 11:00 AM, a second peak at 1:00 PM, and a creative window at 3:30 PM. Her menstrual cycle is in the follicular phase (higher energy, better verbal fluency). The “when to do what” scheduler suggests: client calls at 9:30 AM, admin work during the 11 AM trough, design work during the 1 PM peak, and brainstorming at 3:30 PM. She schedules her week around her biology instead of fighting it.
sa-01-circadian-compass— Ultradian rhythms nest within the circadian frameworksa-09-recovery-score— Cycle phase modulates daily readinesssa-07-botanical-wellness— Phase-specific botanical suggestions (adaptogens during luteal, energizing during follicular)t-01-helio-study— Seasonal daylight changes drive the seasonal energy model
Maps botanical compounds to physiological states with evidence grading. Every suggestion includes: the evidence grade (A = systematic review, B = RCT, C = observational/traditional), known contraindications, and drug interaction warnings from NIH. Slipstream is not a doctor—but it can show you what the research says about chamomile and sleep onset latency.
Current physiological state from sa-02/sa-03/sa-06; concern (sleep, stress, inflammation, focus) (optional)
Botanical suggestions with evidence grade, mechanism, contraindications, drug interactions, dosage ranges from research, and optional XO product linkage (clearly labeled as commercial)
NIH Office of Dietary Supplements (free), PubMed API (free), XO Knowledge Reservoir (internal evidence-graded claims). Commercial layer always opt-in.
Static knowledge base, updated quarterly as new research is reviewed. Physiological matching recalculated daily.
A user’s sleep efficiency has been declining for two weeks. HRV is suppressed. sa-04 shows magnesium intake below DRI. sa-07 surfaces evidence-graded suggestions: magnesium glycinate (Grade A—multiple RCTs for sleep quality), lavender essential oil (Grade B—RCTs for anxiety-related sleep disruption), valerian root (Grade B—modest effect on sleep latency). Each entry links to the PubMed source. Contraindication check: no conflicts with her current medications. The XO Botanicals product line includes a magnesium-lavender blend—shown as a clearly labeled commercial option alongside the informational content.
sa-03-sleep-architecture— Sleep patterns trigger relevant botanical suggestionssa-04-nutrient-compass— Nutritional gaps may be addressable through botanical supplementationsa-06-body-clock— Cycle phase modulates which botanicals are most relevantfg-07-foraging-nutrition— Wild-sourced botanicals available in your area
Not step counting—movement quality, diversity, and preparedness. VO² max is the single strongest predictor of all-cause mortality. But VO² max alone is incomplete. Movement diversity—the variety of planes, loads, and patterns you move through—is the deeper signal. This rill tracks both.
Activity data from sa-02; GPS track logs; optional: Walkability from vi-04 for infrastructure context
Activity classification (walk/run/cycle/swim/lift/yoga), movement diversity score, sedentary bout tracking, VO² max estimation trend with age-contextualized percentile, movement heatmap
Compound: sa-02 for wearable data + device accelerometer/gyroscope + GPS tracks. Pure computation. Free.
Daily. Activity classification happens in near-real-time; diversity and VO² trends update daily.
A 40-year-old runner in Seattle checks sa-08. VO² max estimated at 42 ml/kg/min (72nd percentile for age). But movement diversity is low—she’s been running exclusively for three weeks. The diversity wheel shows no lateral movement, no load-bearing beyond bodyweight, no flexibility work. Stack with c-03 (trail network): “You’ve explored 12% of trails within 5 miles. Here are three hikes with elevation gain that would add vertical and lateral movement to your week.” The movement data meets the map.
c-03-trail-network— Trail difficulty mapped to current fitness for route recommendationscm-03-trail-truth— Real trail conditions matched to your movement capabilitiessa-09-recovery-score— Training load feeds the daily readiness calculationki-04-activity-import— Import GPS tracks from Strava to augment the movement record
A single compound number that answers: “How ready am I today?” Unlike the siloed recovery scores from Whoop, Oura, or Garmin, this one is cross-device and cross-domain—it sees your sleep AND nutrition AND cycle phase AND barometric pressure AND training load. That’s the compound intelligence thesis applied to your own body.
Compound: sa-02 (HRV, resting HR, skin temp), sa-03 (sleep quality/debt), sa-04 (nutrition gaps), sa-06 (cycle phase), sa-08 (training load), a-01 (weather/pressure)
Composite readiness score (0–100), factor breakdown (sleep: 78, recovery: 65, nutrition: 82, cycle: favorable), recommended intensity, specific actionable suggestions, 14-day trend
Compound capstone rill. No external data—synthesizes across the salus family + weather. Free.
Daily. Calculated each morning from overnight data. The first thing you look at.
A triathlete in Portland wakes up to a recovery score of 58. Factor breakdown: sleep 72 (short REM, went to bed late), HRV 61 (below baseline), nutrition 85 (good), training load 44 (heavy week). The recommendation: “Low-intensity day. Zone 2 only. Stretch and hydrate.” She was planning a tempo run. The score doesn’t forbid it—it surfaces the signal. Her own Garmin body battery says 62 (sees only Garmin data). The Slipstream score is lower because it also sees her nutrition gap in magnesium and the high training load from her Strava imports. More data, more honest signal.
sa-08-movement-pattern— Training load is a primary input; the score recommends today’s intensitya-01-hyperlocal-weather— Barometric pressure, temperature, and AQI modulate readinessas-07-horoscope-narrator— For users who compose body data with astrology in their morning briefingco-05-gift-a-flow— A recovery-score morning dashboard is a powerful gift for health-minded friends
Structured health logging with a twist: every entry is auto-tagged with what your body was doing at the time (HRV, sleep debt, cycle phase, weather). Over time, the correlation engine finds patterns you couldn’t see otherwise: “Your headache entries cluster within 8 hours of barometric drops exceeding 6 mb.”
User input: text, tags, sliders (mood, energy, pain), voice memo. Auto-attached: sa-02–sa-09 physiological context at time of entry
Journal feed with physiological sidebar, symptom-trigger correlation analysis (after 30+ entries), intervention tracking, and “did you notice?” pattern alerts
User input + physiological context from salus family + environmental context from a-01, vi-04. Optional V2F (voice-to-feedstock) for voice journaling. Free.
On entry. Correlation engine runs after 30+ entries and updates as the dataset grows.
A user in the SF Bay Area has been logging symptoms for three months. sa-10’s correlation engine surfaces a pattern: “You’ve logged respiratory symptoms 6 times this month. AQI exceeded 100 on 8 of those days. Your journal entries mentioning headaches cluster within 8 hours of barometric drops exceeding 6 mb. You started magnesium supplementation on March 5th—your sleep efficiency has improved 6% since, and headache frequency has decreased.” None of these patterns were visible in the journal alone or the sensor data alone. The correlation emerges because subjective experience meets objective measurement.
vi-04-air-quality— AQI correlated with respiratory symptom journal entriesa-01-hyperlocal-weather— Barometric pressure changes correlated with headache and migraine entriessa-07-botanical-wellness— Intervention tracking validates or contradicts botanical suggestionssa-09-recovery-score— Journal mood entries correlated with recovery score for subjective-objective mapping
vitae (of life)
What is the environment doing to you? Vitae is the environmental health layer—air quality, water quality, toxic proximity, environmental justice screening, and public health outcomes. Where salus tracks what’s happening inside your body, vitae tracks what the environment around you is putting into it. The two families are designed to compose: your respiratory symptoms (sa-10) correlated with the AQI (vi-04) and nearby toxic release sites (vi-03) becomes an environmental health case study.
Ambient air quality and pollutant breakdown, drinking water source and contaminant levels, proximity to Superfund sites and TRI facilities, EPA environmental justice screening, and county-level public health rankings.
EPA AirNow / OpenAQ (air quality, free), EPA SDWIS (drinking water, free), EPA Envirofacts (Superfund, TRI, brownfields, free), EPA EJScreen (environmental justice, free), County Health Rankings (public health, free). All federal, all free, all open.
Air quality alone is a number. Stack vi-04 with cm-09 (citizen sensors) and you reveal where official monitors underrepresent actual exposure. Layer vi-03 (toxic release) with vi-05 (environmental justice) and you surface whether pollution burden disproportionately affects vulnerable demographics. Combine vi-01 (water quality) with cm-05 (community water truth) and you have both the institutional testing data and the community’s lived experience of their water. These compositions are where environmental data becomes environmental advocacy.
Stack vi-04 + vi-03 + vi-05 + cm-09 + sa-10 for a neighborhood in Portland’s inner eastside. Community PurpleAir sensors show PM2.5 readings 35% above the nearest EPA monitor. The toxic release rill identifies two TRI facilities within 2 miles. EJScreen shows the census tract scores in the 85th percentile for pollution burden and 78th for demographic vulnerability. A resident’s health journal (sa-10) has logged respiratory symptoms on 8 of the 12 days when community sensors exceeded AQI 100. The composition surfaces a signal that warrants investigation—the kind of evidence that environmental health researchers and community advocates can use.
| ID | Name | Description |
|---|---|---|
| vi-01 | Water Quality | Drinking water source, violations, and contaminant levels from EPA SDWIS |
| vi-02 | Toxic Release Inventory | Proximity to Superfund sites, TRI facilities, and brownfields from EPA Envirofacts |
| vi-03 | Environmental Justice Screen | EPA EJScreen pollution burden and demographic vulnerability scoring |
| vi-04 | Air Quality | Real-time AQI, pollutant breakdown, and health recommendations from EPA AirNow |
| vi-05 | Public Health Profile | County health rankings, disease prevalence, and healthcare access metrics |
What’s in your drinking water? Source, treatment, contaminant levels, and violation history from the EPA’s Safe Drinking Water Information System. Every public water system in the US is in this database. The question isn’t whether your water has contaminants—it’s which ones, and at what levels relative to the legal limit.
lat, lng or zip_code (required)
Water system ID, source (surface/ground), treatment type, contaminant list with measured levels vs. MCL (Maximum Contaminant Level), violation history, and lead service line probability
EPA SDWIS (Safe Drinking Water Information System). Free. No key required.
Quarterly. EPA violation data updates as systems report. Historical violations persist.
A family considering a home in Tacoma queries vi-01. The water system is City of Tacoma (surface water from the Green River watershed). No current MCL violations. But the historical record shows two lead/copper rule violations in 2019, and the lead service line probability for homes built before 1950 in their target neighborhood is “moderate.” Stack with cm-05 (community water truth) for what residents actually report about their tap water, and with vi-03 to check for contaminated sites upstream.
cm-05-water-truth— Community water observations complement institutional testing datavi-02-toxic-release— Contaminated sites upstream may explain water quality anomaliesh-03-stream-gauge— Source water flow conditions affect treatment demandsvl-01-property-valuation— Water quality violations may affect property values in the area
Proximity to EPA Superfund sites, Toxics Release Inventory facilities, and brownfields. Every year, industrial facilities report what chemicals they release into the air, water, and land. This rill tells you who your neighbors are in the environmental sense—and what they’re putting into the ecosystem.
lat, lng (required); radius (default 5 miles)
Nearby sites: Superfund NPL status, TRI facility names with chemicals released (by type and quantity), brownfield listings, distance and direction from query point
EPA Envirofacts (Superfund, TRI, brownfields). Free. No key required.
Annual. TRI data is reported yearly. Superfund site status updates as remediation progresses.
A couple looking at a house in Portland’s St. Johns neighborhood queries vi-02. Results: the Portland Harbor Superfund site is 1.8 miles south (NPL-listed, active remediation), a TRI facility 2.1 miles east reported 12,000 lbs of volatile organic compounds released in the most recent reporting year, and two brownfield sites within a mile. Stack with vi-03 (environmental justice) and the census tract shows high pollution burden but moderate demographic vulnerability. Layer with vl-01 (property valuation) to assess whether the home price reflects the environmental context.
vi-03-environmental-justice— Toxic proximity overlaid on demographic vulnerability reveals equity patternsvi-04-air-quality— AQI readings near TRI facilities may indicate ongoing exposurecm-09-citizen-sensor— Community sensors near industrial sites provide monitoring that EPA doesn’tvl-01-property-valuation— Environmental contamination should be priced into property values—often it isn’t
EPA EJScreen overlays pollution burden on demographic vulnerability—the tool that answers whether environmental harm is distributed equitably. It isn’t, almost anywhere. This rill makes visible what policy documents bury in appendices: which communities carry disproportionate pollution exposure.
lat, lng (required); radius or census_tract (optional)
EJScreen indices: pollution burden percentile, demographic index percentile, individual indicators (PM2.5, ozone, lead paint, proximity to hazardous waste, traffic density), and comparison to state and national baselines
EPA EJScreen (free). Census ACS demographic data (free). No key required.
Annual. EJScreen dataset updates yearly with new Census and environmental data.
A community organizer in Seattle’s Duwamish Valley queries vi-03. The census tract scores in the 95th percentile nationally for pollution burden and the 82nd percentile for demographic vulnerability. Individual indicators: proximity to hazardous waste (98th percentile), traffic density (91st), and diesel PM (88th). Stack with cm-09 (citizen sensors) showing community PurpleAir readings consistently above the nearest EPA monitor, and the narrative writes itself: a community carrying nearly the highest pollution burden in the state with monitoring infrastructure that undercounts their exposure.
vi-02-toxic-release— TRI data explains why the pollution burden score is highcm-09-citizen-sensor— Community monitoring fills the gaps where official monitoring underrepresents exposurevi-05-public-health— Health outcomes correlated with environmental justice indicescv-08-tax-rate— Tax burden compared to pollution burden reveals fiscal inequity
Real-time AQI with pollutant breakdown and health recommendations. Tells you whether the air outside is safe to breathe, and for whom—because “moderate” air quality means different things for a marathon runner and an asthmatic toddler. Especially relevant during Pacific Northwest wildfire season.
lat, lng (required)
Current AQI, dominant pollutant, pollutant breakdown (PM2.5, PM10, ozone, NO², SO², CO), health category, and sensitivity-group recommendations
EPA AirNow API (free, key required) and/or OpenAQ (free, open-source). PurpleAir community monitors via cm-09 as supplement.
Hourly. AQI updates from official monitors every hour. PurpleAir community sensors update every 2–10 minutes.
In late August, wildfire smoke settles over Portland. vi-04 shows AQI 162 (Unhealthy) with PM2.5 as the dominant pollutant. The health recommendation for sensitive groups: “Avoid prolonged outdoor exertion. Close windows. Run HEPA filter.” Stack with cm-09 (citizen sensors) and PurpleAir monitors on the east side show AQI 195, while the west hills show AQI 140—terrain channeling the smoke. Layer with sa-10 (health journal) to track whether symptoms correlate with the smoke event. The composition turns a generic AQI number into a personal health decision surface.
cm-09-citizen-sensor— PurpleAir monitors reveal hyperlocal variations the official network missessa-10-health-journal— Respiratory symptom entries correlated with AQI spikessa-08-movement-pattern— AQI should modulate outdoor exercise recommendationsa-01-hyperlocal-weather— Wind direction and inversion layers explain why AQI varies within a city
County-level health rankings, disease prevalence, healthcare access, and health behavior data. The population-level context for individual health data—because your personal health exists within a community health landscape, and knowing the landscape changes what you pay attention to.
lat, lng or county_fips (required)
County health ranking, premature death rate, adult obesity rate, physical inactivity rate, primary care physician ratio, uninsured rate, and comparison to state medians
County Health Rankings (University of Wisconsin, free). CDC WONDER (free). Census ACS for demographic context. Free.
Annual. County Health Rankings update each March with prior-year data.
A family evaluating a move to rural Clatsop County on the Oregon coast queries vi-05. The county ranks 22nd of 36 Oregon counties in health outcomes. Key concerns: primary care physician ratio of 2,100:1 (state median: 1,260:1) and the nearest hospital is 45 minutes away. Physical inactivity rate is below state median (good). Stack with vi-04 (air quality—excellent on the coast) and vi-01 (water quality—clean aquifer). The health profile reveals a tradeoff: great environmental health, limited healthcare access. That’s a decision surface for a family with specific health needs.
vi-03-environmental-justice— Pollution burden correlated with health outcomes at the county levelsa-09-recovery-score— Personal health in the context of community health infrastructurevl-08-cost-of-living— Health outcomes alongside economic indicators for relocation decisionscv-06-school-district— School quality and health outcomes compose a family livability picture
astra (stars, celestial bodies)
The Cosmo family handles astronomy—celestial mechanics, orbital positions, measurable phenomena. Astra handles astrology—the interpretive framework humans have layered on top of those mechanics for 4,000 years. The key insight: astrology is computed from the same data as astronomy. Planetary positions are real. Aspects (angular relationships between planets) are real geometry. The interpretation is the cultural and spiritual layer. Slipstream computes the positions with precision and presents the interpretive framework without claiming scientific validity—the same way a museum might exhibit ancient astrolabes alongside modern GPS.
Natal chart computation (sun/moon/rising/all planets, houses, aspects), daily transits relative to birth chart, lunar sign and void-of-course tracking, retrograde monitoring, relationship synastry, seasonal astrology (ingresses, eclipses), personalized horoscope narration, planetary hours, solar return charts, and lunar node positions.
Swiss Ephemeris via kerykeion or flatlib Python libraries (free, open-source, gold standard for astrological computation). No external API needed—all computation is local from bundled ephemeris data. Claude API for narrative generation (as-07). Zero cost for computation rills, zero rate limits, works offline.
Astrology alone is a personal reflection framework. Stack as-03 (Moon Tracker) with sp-02 (Sports Pulse) and a-01 (weather) and you have a morning briefing someone opens every day. Layer as-01 (Birth Chart) with sa-01 (Circadian Compass) and the astronomical positions that astrology interprets overlap with the solar geometry that drives your body clock. Compose as-07 (Horoscope Narrator) with sa-09 (Recovery Score) and the daily reading arrives alongside your body’s readiness data—two different languages describing the same morning. For users who engage with both frameworks, the composition is deeply personal.
Stack as-03 + as-02 + as-07 + sa-09 + sp-02 + a-01 for a complete morning briefing. Moon in Scorpio—emotional depth, intensity. Transiting Mars squares natal Venus at 0.8° orb: “Tension between what you want and what you need may surface today. An invitation to examine desires with honesty.” Recovery score: 78. Warriors play tonight at 7:30 PM. Weather: clear, 58°F, golden hour at 5:18 PM looks exceptional. Planetary hour: Jupiter rules the first hour after sunrise—traditionally favorable for expansion and generosity. Each layer adds a lens. The whole is a morning ritual that combines body, sky, culture, and place.
| ID | Name | Description |
|---|---|---|
| as-01 | Birth Chart | Full natal chart from birth date, time, and location via Swiss Ephemeris |
| as-02 | Daily Transit | Current planetary positions aspecting your natal chart—the real horoscope |
| as-03 | Moon Tracker | Current moon sign, phase, void-of-course windows, and lunar aspects |
| as-04 | Retrograde Monitor | Mercury and all planetary retrogrades with station dates and practical timing |
| as-05 | Synastry | Relationship chart comparison between two natal charts with composite generation |
| as-06 | Seasonal Astrology | Ingresses, eclipses, and major aspect patterns for the current season |
| as-07 | Horoscope Narrator | Claude-powered personalized daily reading from your actual transit data |
| as-08 | Planetary Hours | Traditional planetary hour calculator for the current day and location |
| as-09 | Solar Return | Annual chart cast for the exact moment the sun returns to your natal position |
| as-10 | Lunar Nodes | North and south node positions, eclipse axis, and nodal return tracking |
Computes a full natal chart from birth date, time, and location—sun sign, moon sign, rising sign, all planetary placements, house positions, and major aspects. The chart wheel rendered as SVG is the foundation for everything else in the astra family. Think vintage astronomical illustration meets clean modern design.
birth_date (required); birth_time (optional—noon used if missing, houses approximate); birth_location (lat/lng, required for rising sign); house_system (Placidus default, Whole Sign, Koch, Equal)
Natal chart wheel (SVG), planet placement table (sign, degree, house, dignity), aspect grid, summary (sun/moon/rising, dominant element and modality), and raw ephemeris data
Swiss Ephemeris via kerykeion (Python, open-source). Bundled ephemeris data. No API, no key, no cost. Works offline.
Static. A natal chart is computed once from birth data. Recalculated only if birth time is updated.
For a birthday gift, someone enters their partner’s birth data: February 14, 1990, 3:22 AM, San Francisco. as-01 computes the chart: Sun in Aquarius (humanitarian, unconventional), Moon in Cancer (emotional depth, nurturing), Rising in Sagittarius (adventurous, philosophical). Dominant element: water (4 planets). The SVG chart wheel is beautiful enough to frame. The aspect grid reveals a tight Venus-Mars conjunction in Capricorn—the astrologer’s interpretation unfolds from there. Stack with as-05 (synastry) using both partners’ charts for the relationship overlay.
as-02-daily-transit— Transits are computed relative to the natal chart—this is the foundationas-05-synastry— Two birth charts overlaid reveal relationship dynamicsas-09-solar-return— Annual chart calculated from the natal sun’s exact positionco-05-gift-a-flow— A natal chart + daily transit flow is a birthday gift that keeps giving all year
Current planetary positions relative to your natal chart—what’s “happening” astrologically today. Not a generic sun-sign horoscope but transiting planets aspecting your natal planets, computed from your actual birth chart. The real horoscope is personal; everything else is a newspaper column.
Natal chart from as-01 (required); date (default: today); orb_threshold (default: 2° for personal planets, 1° for outer)
Transit-to-natal aspect list with exact times, orb, applying/separating status, and archetypal keyword summary. Significant transits highlighted.
Swiss Ephemeris for current positions + natal chart from as-01. Pure computation. No API, no key, free.
Daily. Recalculated each morning. Fast-moving transits (Moon, Mercury) update through the day.
A user opens as-02 and sees: transiting Jupiter trines natal Moon (1.2° orb, applying)—traditionally interpreted as emotional expansion, generosity, a good day for family. Transiting Mars squares natal Venus (0.8°, exact tomorrow)—tension between desire and relationship, the energy needs an outlet. The orb and applying/separating status tell you not just what’s happening but whether it’s building or fading. Stack with as-07 for Claude’s narrative interpretation of these specific transits.
as-07-horoscope-narrator— Claude wraps the raw transit data in warm, personalized proseas-03-moon-tracker— The fastest transit, updated throughout the daysa-09-recovery-score— Body data and transit data side by side in the morning briefingas-01-birth-chart— Transits only make sense relative to the natal chart they aspect
The moon changes sign every ~2.5 days—the fastest-moving astrological body and the one most astrology-interested people track daily. Current sign, phase, void-of-course windows, and today’s lunar aspects. The rill someone checks every morning before checking the weather.
lat, lng (for local timing of moonrise/set); date (default: today)
Current moon sign and degree, phase with illumination %, void-of-course status (with start/end times), next 7 days of sign changes, today’s lunar aspects with exact times, element and modality
Swiss Ephemeris computation. Same math as h-02 (Lunar Phase) but with the astrological interpretation layer. No API, free.
Updates throughout the day as the moon moves. Sign changes and void-of-course windows calculated for the full week ahead.
Devin’s mom opens her morning digest. Moon in Libra—harmony, balance, relationships. Void of course from 2:15 PM to 6:30 PM (in traditional astrology, a window to avoid starting new things). Full Moon in Virgo at 9:17 PM—the exact opposition to her natal Neptune suggests heightened intuition and vivid dreams tonight. The moon tracker is the daily driver of the astra family: fast-moving, always relevant, and deeply personal when layered on a natal chart.
sp-02-sports-pulse— Moon sign + game alerts = the complete morning briefingsa-01-circadian-compass— Lunar cycle overlaid on circadian rhythm for dual-framework trackingas-02-daily-transit— Lunar aspects are the fastest-changing transits in the daily pictureh-02-lunar-phase— Same astronomical data, different interpretation: tides vs. emotional weather
Which planets are currently retrograde, when did they station, and when do they go direct? Mercury retrograde is the famous one, but all planets except the sun and moon have retrograde periods. This rill tracks them all and includes the pre-retrograde shadow period that astrologers often reference.
None required (current date used). Optional: date_range for upcoming retrogrades
Retrograde status dashboard: which planets are currently Rx, station dates (retrograde and direct), degrees at station, shadow period dates, and upcoming retrograde timeline for the year
Swiss Ephemeris. Retrograde = when ecliptic longitude is decreasing (apparent backward motion). Pure computation, free.
Daily. Retrogrades last weeks to months. Station dates (the day a planet appears to stop) are the key events.
Mercury stations retrograde tomorrow. as-04 alerts: “Mercury Rx begins March 15 at 9° Aries, goes direct April 7 at 26° Pisces. Shadow period began March 1; post-shadow clears April 21.” For users who track retrogrades as a timing framework, this is practical calendar information. Saturn is also retrograde, and Mars enters its shadow period next week. The annual timeline view shows all upcoming retrogrades for planning purposes—whether for practical caution or as a reflective framework.
as-02-daily-transit— Retrograde transits to natal planets have distinct interpretive weightas-07-horoscope-narrator— Narrative context for retrograde periods in the daily readingas-06-seasonal-astrology— Retrogrades are part of the seasonal astrological weather
Overlay two birth charts and see where they connect—harmonious aspects, tension points, and the composite chart that represents the relationship as its own entity. Not a compatibility score; a map of the dynamic between two people, rendered with the same geometric precision as the individual charts.
Two natal charts from as-01 (birth data for both people)
Dual-wheel overlay (SVG), inter-chart aspect grid, composite midpoint chart, highlighted harmonious and challenging aspects, and element/modality compatibility analysis
Swiss Ephemeris computation from two birth datasets. No external API. Free.
Static. Synastry is computed from fixed birth data. Recalculated only if birth times are updated.
A couple curious about their chart dynamics enters both birth data into as-05. The synastry reveals: her Venus conjuncts his Mars (traditionally strong attraction), but her Saturn squares his Moon (a pattern where emotional expression meets caution). The composite chart shows a Sun-Jupiter conjunction in the 7th house—a traditionally favorable placement for partnership. The dual-wheel SVG shows both charts overlaid, with colored aspect lines connecting the two. It’s not a compatibility percentage—it’s a map of the relational terrain.
as-01-birth-chart— Two individual charts are the inputas-02-daily-transit— Current transits to the composite chart show the relationship’s weatherco-05-gift-a-flow— A synastry chart is an intimate, unique gift
The big astrological weather that affects the collective—ingresses (planets changing signs), eclipses, and major aspect patterns (grand trines, T-squares, grand crosses). Not personalized to your chart, but the seasonal backdrop against which your transits play out.
season or date_range (optional; defaults to current season and 3 months forward)
Seasonal calendar: planet ingress dates, eclipse dates and types, major multi-planet patterns, retrograde stations, and solstice/equinox markers
Swiss Ephemeris for ingress/eclipse computation. Same data as cs-05 (celestial events) but with astrological interpretation. Free.
Seasonal. Major patterns computed for the quarter ahead. Eclipse data precise to the minute.
Opening as-06 for spring 2026 shows: Aries ingress (spring equinox) March 20, a solar eclipse in Aries on March 29, Jupiter ingresses Cancer on April 11 (traditionally its exaltation sign), and a Saturn-Neptune conjunction at 0° Aries in June—an aspect that occurs roughly every 36 years. The seasonal view provides context for individual daily transits: “Your Mars return happens during the Saturn-Neptune conjunction—bigger context for a personal transit.”
as-02-daily-transit— Personal transits gain meaning against the seasonal backdropag-02-seasonal-calendar— Astrological seasons overlay the natural-cycle calendaras-04-retrograde-monitor— Retrogrades are major events in the seasonal astrological picture
Claude-powered personalized daily reading based on your actual transits, not your sun sign. Warm, reflective, empowering—never fatalistic. Treats astrology as a framework for self-reflection, not prediction. References specific transits by name and explains the archetypal meaning in plain language.
Transit data from as-02; natal chart from as-01; frequency (daily or weekly); depth (brief: 3–5 sentences, full: 2–3 paragraphs)
Prose narrative: personalized interpretation of today’s most significant transits, archetypal themes, and reflective prompts. Always uses “may,” “could,” “an invitation to”—never “will” or “must.”
Transit data from as-02 + Claude API with astrology-literate system prompt. Claude API cost per reading.
Daily (or weekly). Generated fresh each morning based on that day’s transits.
Mom’s morning digest includes today’s narrative: “Transiting Jupiter trines your natal Moon today at 1.2°—an expansive, generous emotional energy. This is a day that favors connection, family, and starting things that bring you joy. Meanwhile, Mars approaching a square to your Venus this week may surface tension between what you want and what feels comfortable. An invitation to examine desires with honesty rather than judgment.” The narrative is warm but specific—it references her actual chart, not generic Pisces predictions.
as-02-daily-transit— The raw transit data that the narrator interpretssp-02-sports-pulse— Side by side in the morning digest: soul and sportssa-09-recovery-score— Body data alongside cosmic narrative—two lenses on the same morningco-05-gift-a-flow— A daily personalized horoscope is the gift that keeps giving
An ancient timing system that divides daylight and nighttime into 12 unequal “hours,” each ruled by a classical planet. The concept predates mechanical clocks—an hour of Saturn has a different quality than an hour of Jupiter. Computed precisely from sunrise/sunset at your location.
lat, lng (required); date (default: today)
Today’s planetary hour table: start/end times for all 24 planetary hours (12 day, 12 night), ruling planet for each, current hour highlighted, day ruler (the planet that rules the day of the week)
Sunrise/sunset computed from solar ephemeris + latitude. Planetary hour sequence is fixed by tradition. Pure computation, free.
Daily. Planetary hours change length as day length changes with the seasons. Recalculated each day.
A practitioner in Portland opens as-08 on a Wednesday (Mercury’s day). Sunrise at 6:48 AM, sunset at 5:52 PM. Each daytime planetary hour is ~55 minutes (shorter in winter). The first hour after sunrise is ruled by Mercury (the day ruler). The current hour (2:00 PM) is ruled by Jupiter—traditionally associated with expansion, teaching, and generosity. For users who work with planetary hours as a timing framework, this provides precise, location-specific scheduling guidance.
t-01-helio-study— Sunrise/sunset times are the foundation for planetary hour calculationas-02-daily-transit— Planetary hours add a temporal layer to the transit picturesa-06-body-clock— Ultradian rhythms can be compared to planetary hour rhythms for the curious
A chart cast for the exact moment the sun returns to its natal position each year—your astrological “birthday chart.” Traditionally used to forecast the themes and energy of the year ahead. Because the return happens at a slightly different time each year, the rising sign and house placements shift, creating a unique annual snapshot.
Natal chart from as-01; return_year (default: current); location (where you’ll be on your birthday—matters for house placements)
Solar return chart wheel (SVG), planet placements in return houses, comparison to natal chart, highlighted themes based on house emphasis, and year-ahead aspect timeline
Swiss Ephemeris. The exact return moment computed to the minute. Free, no API.
Annual. Computed once per birthday year. Updated if birthday location changes.
A user’s birthday approaches. as-09 computes the solar return chart for this year: the sun returns to 24° Aquarius at 3:47 AM on February 13 at her current location in Portland. The return chart shows Sagittarius rising (a year of exploration and learning), with Mars in the 10th house (career energy, visible ambition) and Venus conjunct Jupiter in the 5th house (creativity, joy, possibly romance). Compare to last year’s return chart, which had Saturn on the Ascendant—a harder, more disciplined year. The shift in energy is visible in the chart.
as-01-birth-chart— Solar return is interpreted relative to the natal chartas-02-daily-transit— Transits to the solar return chart add another timing layerco-05-gift-a-flow— A solar return reading is a meaningful birthday gift
The lunar nodes are the points where the moon’s orbit crosses the ecliptic—the astronomical reality behind eclipses. In astrology, the north node represents growth direction and the south node represents what you’re releasing. The nodes change sign every ~18 months, and their transits through your chart mark significant life chapters.
Optional: natal chart from as-01 for personalized node analysis. Without it, shows current node positions and eclipse axis.
Current north/south node positions (sign and degree), eclipse axis, upcoming eclipses on the nodal axis, natal node positions, transiting nodes aspecting natal planets, and nodal return timing (~18.6 year cycle)
Swiss Ephemeris. Lunar node computation from orbital mechanics. Free, no API.
Monthly. The nodes move slowly (~1.5° per month, retrograde). Eclipse dates are the key trigger events.
A user in their late 30s checks as-10 and discovers they’re approaching a nodal return—the transiting nodes returning to their birth position, which happens at ages ~19, ~37, ~56. In astrology, this marks a significant recalibration of life direction. The current eclipse axis (Aries-Libra) is aspecting her natal Moon. An upcoming lunar eclipse falls within 2° of her natal Moon—a transit that astrologers consider particularly significant for emotional and domestic themes. The rill doesn’t predict what will happen; it maps the geometry and lets her interpret.
as-06-seasonal-astrology— Eclipses on the nodal axis are the most significant seasonal eventsas-02-daily-transit— Nodal transits to natal planets appear in the daily pictureas-01-birth-chart— Natal node positions reveal the karmic axis in the birth chart
Fabrica — Materials & Making
Everything built so far flows in one direction: world → sense → interpret → person. Content rills observe. Multiplier rills contextualize. Community rills verify. The person receives intelligence. Fabrica reverses the current. The person has an intention—grow something, build something, make something, express something—and the assembled rills inform that intention with everything they know about the place. The creation rills generate the output: a glaze formula, a toolpath, a soundscape, a projection map, a festival score.
These rills draw from USGS geological surveys, SSURGO soil mineralogy, Glazy.org glaze databases, Swiss Ephemeris, TouchDesigner protocols, WLED/ArtNet/DMX lighting standards, Web Audio APIs, and the full stack of content rills that feed creative decisions with site-specific intelligence. Individually, a clay mineral profile is a geology curiosity. Composed with glaze chemistry, kiln regulations, and a palette extracted from the autumn canopy overhead, it becomes the complete material intelligence for a potter who just arrived and wants to know: what can I make from what’s here?
materia (material, substance, timber)
Every place has materials. Clay under the soil. Stone in the hillside. Wood in the forest. Fiber in the plants. Pigment in the earth. Materia rills answer: what can you make from what’s here? This is the knowledge that potters, builders, dyers, herbalists, and makers have carried for millennia—largely as oral tradition, now fragmenting as craft communities thin. Materia digitizes bioregional material intelligence so that a ceramicist, a natural dyer, a furniture maker, or a fermentation enthusiast can arrive at a new place and immediately understand its creative potential.
Clay mineralogy and firing characteristics (kaolinite, illite, smectite, montmorillonite), ceramic glaze chemistry in UMF notation, earth pigment deposits (ochres, siennas, umbers, copper carbonates), stone and timber properties (hardness, grain, rot resistance, workability), natural fiber and dye plant ranges, botanical formulation for aromatics and cosmetics, fermentation terroir (wild yeast, water chemistry, local substrate), and secondary materials from salvage and waste streams.
USGS geological maps and mineral deposit databases (free). SSURGO soil survey clay fraction and Munsell color data (free, USDA). Glazy.org ceramic chemistry database (free, community). USDA PLANTS database and USFS Forest Inventory & Analysis (free). Ethnobotanical databases and dye plant range maps. Phytochemistry databases for essential oil yield and active compounds. Building permit demolition data and salvage yard networks.
Materia alone tells you what’s in the ground. Stack mt-01 (Local Clay) with mt-02 (Glaze Chemistry) and the geology rills, and a potter arriving at a new location immediately knows where to dig, what the clay will do in the kiln, and which local minerals yield glaze recipes. Layer mt-06 (Botanical Formulation) with flora rills and the XO Botanicals product line gets a site-specific ingredient intelligence engine. Compose mt-07 (Fermentation Terroir) with water quality and seasonal harvest data and a brewer, cheesemaker, or sourdough baker understands the invisible biology of place that shapes flavor. Every materia rill transforms “I wonder what I could make here” into “here’s exactly what this place offers.”
Stack mt-01 + mt-02 + mt-03 + geology rills + water quality + community knowledge for a complete ceramic material intelligence flow. The geology says: Miocene basalt with weathered feldspar. The soil survey says: 28% clay fraction, high in illite with iron oxide inclusions. mt-01 translates: this clay will fire to a warm terra cotta at cone 06, with good plasticity and moderate shrinkage. mt-02 computes: local feldspar + limestone + oak wood ash yields 14 possible glaze recipes in UMF notation, three in the celadon family. mt-03 adds: iron oxide deposits 2.3 km southeast yield a reliable ochre for slip decoration. Community knowledge confirms: “There’s good red clay in the cut bank by the mill road.” Confidence: 0.92. A potter’s complete material survey, assembled from federal data and community intelligence.
| ID | Name | Description |
|---|---|---|
| mt-01 | Local Clay | Clay minerals in nearby soil and geology—kaolinite, illite, smectite, montmorillonite—with predicted firing characteristics |
| mt-02 | Glaze Chemistry | Ceramic glaze recipes computed from local mineral sources: feldspar, limestone, iron ochre, wood ash in UMF notation |
| mt-03 | Natural Pigment | Earth pigment deposits nearby: iron oxides, manganese blacks, copper carbonates, chalk whites mapped from geology |
| mt-04 | Stone & Timber | Local buildable materials: stone type, timber species, hardness, grain, rot resistance, and workability properties |
| mt-05 | Fiber & Dye | Natural fibers and dye plants in range: cotton, flax, hemp, indigo, madder, weld, walnut hull, osage orange |
| mt-06 | Botanical Formulation | Aromatic, medicinal, and cosmetic-grade botanicals with essential oil yields and active compound profiles |
| mt-07 | Fermentation Terroir | Location-specific fermentation potential: wild yeast populations, water mineral content, local substrate availability |
| mt-08 | Salvage & Reuse | Secondary materials nearby: architectural salvage, reclaimed lumber, industrial surplus, and recoverable waste streams |
What clay minerals exist in the soil and geology near you? Predicts clay body characteristics from geological survey data—kaolinite for porcelain-like whiteness, illite for workable stoneware, smectite for high plasticity, montmorillonite for dramatic shrinkage. Wild clay potters spend years learning to read landscape for clay. Geological data tells you where to dig and what you’ll find.
Latitude/longitude or address. Optional: target firing temperature, specific mineral interest.
Clay mineral composition within search radius, predicted firing range (earthenware/stoneware/porcelain temperatures), plasticity estimate, shrinkage percentage, suggested dig locations ranked by accessibility, and Munsell color predictions for raw and fired states.
USGS geological maps (free). SSURGO soil survey clay fraction data (free, USDA). Mineralogy databases cross-referenced with geological formation age and parent material.
Static—geological data updates on decadal survey cycles. Cache indefinitely for a given location.
Given local mineral sources—feldspar, calcium from limestone, iron from ochre, wood ash composition from local trees—compute possible glaze recipes. Maps geological minerals to ceramic oxide chemistry using Unity Molecular Formula analysis. Glaze chemistry is graduate-level ceramics knowledge; the math is straightforward but the access to the knowledge is the gate. Connects directly to Le Four Argile.
Location (for local mineral sourcing) or manual mineral list. Optional: target surface (matte, satin, gloss), color family, firing cone.
Glaze recipes in UMF notation with local mineral proportions, predicted surface quality at target temperature, thermal expansion coefficient for clay body compatibility, and links to similar recipes in the Glazy.org community database.
mt-01 (Local Clay) + geology rills for mineral identification. Glazy.org database for recipe cross-referencing (free, community). Wood ash chemistry tables keyed to local tree species.
Static mineral data, community database updates weekly. Recompute when new mineral sources are identified.
What earth pigments exist nearby? Maps geological deposits to pigment potential: iron oxides yield ochres, siennas, and umbers. Manganese deposits yield blacks. Copper carbonates yield greens and blues. Chalk yields whites. Artists have been grinding local pigments for 40,000 years. The geological data to find them exists; the translation to art-material doesn’t—until now.
Latitude/longitude or address. Optional: target color family, medium (oil, watercolor, fresco, ceramic slip).
Pigment-grade mineral deposits within search radius, predicted color in Munsell notation, collection accessibility, processing requirements (grinding, washing, levigation), and historical use context for the deposit type.
USGS mineral deposit databases (free). Soil color data from SSURGO Munsell measurements (free). Geological survey cross-referenced with known pigment mineral associations.
Static geological data. Cache indefinitely for a given location.
What buildable materials exist locally? Stone type—granite, sandstone, limestone, basalt—with hardness, porosity, and weathering characteristics. Timber species with properties: Janka hardness, grain pattern, rot resistance, workability, and traditional uses. A builder or furniture maker traditionally learns local materials through apprenticeship. The forestry and geology data exists federally; the craft translation is what’s been missing.
Latitude/longitude or address. Optional: intended use (construction, furniture, carving, fencing), scale of project.
Dominant stone types with structural and aesthetic properties. Timber species inventory with Janka hardness, specific gravity, grain description, dimensional stability, and workability ratings. Suggested applications for each material based on its properties.
USGS geology maps (free). USFS Forest Inventory & Analysis (FIA) for timber species and volume (free). Wood properties databases (USDA Forest Products Laboratory). Quarry and sawmill location data.
Geology is static. Forest inventory updates on 5–10 year survey cycles. Timber market availability is more dynamic.
What natural fibers grow here—cotton, flax, hemp, wool-bearing animals? What natural dyes can be extracted from local plants—indigo, madder, weld, cochineal host plants, walnut hull, osage orange? Fiber arts and natural dyeing are experiencing a revival, but knowledge is scattered across books, workshops, and oral tradition. This rill assembles dye plant range maps, fiber crop suitability, and traditional extraction methods into a single bioregional textile intelligence layer.
Latitude/longitude or address. Optional: target fiber type, target color, seasonal availability window.
Fiber crops suited to the location’s climate and soil. Dye plants in range with color yield, mordant requirements, harvest timing, and traditional preparation methods. Seasonal availability calendar for foraged dye materials. Lightfastness and washfastness ratings for each dye source.
Flora rills (f-01, f-02) for plant range data. Ethnobotanical databases for dye plant associations. USDA PLANTS database for fiber crop suitability. Dye plant range maps from textile conservation literature.
Plant range data is seasonally stable. Foraged dye availability shifts with phenology—refresh monthly during growing season.
What aromatic, medicinal, or cosmetic-grade botanicals grow or can be cultivated at this location? Essential oil yield estimates, carrier oil sources, active compound profiles. Formulation chemistry is the gatekept knowledge—knowing which local plants yield which compounds at what concentrations. Connects directly to XO Botanicals, making site-specific ingredient intelligence available for product development from raw landscape data.
Latitude/longitude or address. Optional: target compound class (essential oils, carrier oils, hydrosols, waxes), application (skincare, fragrance, therapeutic).
Botanical inventory with essential oil yield percentages, dominant chemical constituents (linalool, geraniol, eugenol, etc.), carrier oil fatty acid profiles, climate suitability scores for cultivation, and seasonal harvest windows for peak compound concentration.
Flora rills for species range data. USDA plant database for cultivation parameters. Phytochemistry databases for compound profiles. Climate suitability models for aromatic crop viability. Essential oil yield literature from distillation studies.
Species range data is seasonally stable. Compound profiles are static per species. Cultivation suitability may shift with climate projections.
What can you ferment here? Wild yeast and bacteria populations are location-specific. Water mineral content affects brewing. Local fruit, grain, and honey availability determines substrate. The invisible biology of place shapes flavor—every great fermentation tradition (wine, beer, cheese, sourdough, miso) is rooted in local microbiome and materials. The terroir concept, generalized beyond wine into a universal fermentation intelligence layer.
Latitude/longitude or address. Optional: fermentation type (beer, wine, cider, cheese, sourdough, miso, kombucha), target flavor profile.
Water mineral profile and its effect on fermentation (hardness, pH, mineral ratios). Local fruit, grain, and honey availability by season. Wild yeast capture probability and seasonal timing. Fermentation tradition context for the region. Suggested starter projects based on available substrates.
Water quality data (v-01 if available). Flora rills for fruit and grain range. Soil microbiome data (emerging). Regional fermentation tradition databases. Seasonal harvest calendar from forage rills.
Water chemistry is relatively stable. Substrate availability is seasonal—refresh monthly during harvest season. Microbiome data is emerging and updates as research publishes.
What secondary materials are available nearby? Architectural salvage, reclaimed lumber, industrial surplus, and waste streams that become art materials. Construction demolition patterns, manufacturing waste, municipal material recovery. The circular economy as material intelligence—what’s being torn down near you? What’s being thrown away that you could make with? This rill maps the invisible material flows of demolition, surplus, and recovery.
Latitude/longitude or address. Optional: material type interest (lumber, metal, glass, stone, fixtures), project scale, search radius.
Active demolition permits within radius (signaling incoming salvage material). Salvage yards and architectural reclamation centers. Industrial surplus outlets. Materials exchange networks (Freecycle, Buy Nothing, Habitat ReStore). Waste stream characterization for the area’s dominant industries.
Building permit data (demolition permits from municipal open data). Salvage yard and ReStore location databases. Freecycle/materials exchange network APIs. Industrial ecology databases for regional waste stream characterization.
Demolition permits update weekly. Salvage yard inventory is dynamic—daily for active yards. Materials exchange posts refresh hourly.
fabrica (workshop, craft, art)
Fabrica rills are the output layer—they take creative decisions informed by all the other families and produce files, signals, and instructions that drive physical and digital fabrication. These are the “last mile” rills that bridge between Slipstream’s intelligence and the tools in your studio, your workshop, your codebase. TouchDesigner, WLED, pen plotters, CNC mills, laser cutters, MIDI controllers, projection mapping rigs, PCB fabrication houses, and print-ready PDF—fabrica makes data leave the screen and enter the physical world.
TouchDesigner-compatible data export (CHOP/TOP/SOP/DAT), GLSL/WGSL shader generation, LED pixel mapping for WLED and ArtNet, toolpath optimization for pen plotters and laser cutters, CNC and 3D print file generation from terrain data, MIDI/OSC real-time data routing, publication-ready print composition, projection mapping content with surface warping, PCB layout generation, and workshop equipment monitoring.
Any upstream rill with numerical, spatial, or time-series output. TouchDesigner protocols (OSC, WebSocket). WLED/ArtNet/sACN lighting protocols. WebMIDI and Ableton Link for audio routing. SVG/DXF/G-code standards for CNC and laser. STL/3MF for 3D printing. KiCad scripting for PCB design. Syphon/NDI/Spout for projection. PDF/InDesign for print.
Fabrica is what makes Slipstream not just a screen-based system but a tool that drives physical creative practice. Stack fb-01 (TouchDesigner Bridge) with any data rill and you have a live installation feed. Layer fb-03 (LED Mapper) with po-03 (Light Scribe) and bird migration density drives an LED wall in real time. Compose fb-04 (Plotter & Laser) with cartographic layers and your watershed becomes a pen-plotted print on your wall. fb-07 (Print Composer) assembles any multi-rill composition into a museum-quality broadsheet. The data flows through the screen and out the other side—into LEDs, projectors, plotters, CNC mills, speakers, and printed matter.
Stack fb-01 + fb-03 + fb-06 + fb-08 + poiesis rills for a complete responsive installation pipeline. Real-time weather data flows through po-01 (Data Painter) into visual parameters, exported via fb-01 to TouchDesigner. Simultaneously, po-02 (Sonifier) maps the same data to audio, routed through fb-06 as MIDI CC values to a synthesizer. fb-03 maps po-03’s LED patterns to a 400-pixel addressable strip via WLED. fb-08 generates projection content for two synchronized projectors with edge blending. The result: a gallery installation where wind speed, bird density, and soil moisture from a specific watershed drive light, sound, and projection in real time. The place performing itself.
| ID | Name | Description |
|---|---|---|
| fb-01 | TouchDesigner Bridge | Exports any rill’s data as TouchDesigner-compatible formats: CHOP channels, TOP textures, SOP geometry, DAT tables |
| fb-02 | Shader Generator | Translates visual parameters into GLSL/WGSL fragment shaders for WebGL, TouchDesigner, or shader editors |
| fb-03 | LED Mapper | Maps Light Scribe output to physical LED hardware: WLED segments, ArtNet universes, addressable pixel maps |
| fb-04 | Plotter & Laser | Converts SVG-generating rills into optimized toolpaths for pen plotters, laser cutters, and vinyl cutters |
| fb-05 | Topo Carver | Converts terrain data into CNC toolpaths or 3D-printable models—your watershed carved in wood |
| fb-06 | MIDI/OSC Emitter | Real-time rill data routed as MIDI CC values, MIDI notes, or OSC messages to synthesizers and DAWs |
| fb-07 | Print Composer | Assembles multi-rill outputs into publication-ready layouts: poster, broadsheet, field guide, zine spread |
| fb-08 | Projection Content | Generates projection-mapping-ready video with surface warping, edge blending, and multi-projector support |
| fb-09 | PCB Designer | Circuit board layout generation and routing from schematics—sensor boards, LED drivers, custom controllers |
| fb-10 | Machine Monitor | Real-time status and maintenance tracking for workshop equipment: 3D printers, CNC, kilns, laser cutters |
Exports any rill’s data as TouchDesigner-compatible formats: CHOP channels for time-series data, TOP textures for spatial data, SOP geometry for 3D terrain, DAT tables for structured records. Live OSC/WebSocket feed for real-time installations. This is the primary bridge between Slipstream’s data intelligence and the visual programming environment where installations, performances, and responsive environments are built.
Any rill output (numerical, spatial, time-series, tabular). Configuration: target TouchDesigner operator type, update rate, data mapping.
.tox component files for drag-and-drop import. Live OSC streams at configurable port and address pattern. WebSocket JSON feed for web-based TD projects. CSV export formatted for DAT import. CHOP channel mapping documentation.
Any upstream rill. TouchDesigner OSC protocol specification. WebSocket standard (RFC 6455). Data format adapters for each rill output type.
Real-time for live installations (sub-second latency via OSC/WebSocket). Batch export for pre-rendered content. Rate configurable per use case.
Translates visual parameters from Data Painter or Palette Extractor into GLSL/WGSL fragment shaders. Data-driven shader code you can drop into a WebGL scene, TouchDesigner GLSL TOP, or standalone shader editor. The parameters are mapped to uniforms tied to rill data, so the shader responds to real environmental conditions—soil moisture drives displacement, wind speed drives turbulence, temperature gradient drives color ramp.
Visual parameters from po-01 (Data Painter) or fm-01 (Palette Extractor). Target platform: WebGL, TouchDesigner GLSL TOP, Shadertoy, or raw GLSL/WGSL.
.glsl / .frag / .wgsl files with parameterized uniforms mapped to rill data sources. Includes uniform declaration block, main fragment function, and documentation of which rill feeds which visual parameter.
po-01 (Data Painter) visual parameter mappings. fm-01 (Palette Extractor) color values. Any numerical rill data for uniform binding. GLSL/WGSL language specifications.
Shader code is generated once per configuration change. Uniform values update in real time from data feeds.
Takes Light Scribe output and maps it to specific LED hardware configurations: WLED segments with pixel-level addressing, ArtNet universes for professional DMX fixtures, pixel maps for addressable strips and matrices. Configure your physical LED layout—strip length, pixel density, arrangement geometry—and Slipstream drives the pattern. Bird migration density pulsing through a ceiling installation. Soil moisture gradients rendered across a wall of LEDs.
LED pattern data from po-03 (Light Scribe). Hardware configuration: pixel count, arrangement (linear, matrix, custom map), protocol (WLED, ArtNet, sACN).
WLED preset JSON for direct upload. ArtNet DMX universe data stream. Pixel map configuration file describing physical-to-logical mapping. Preview visualization of the mapped pattern on the configured layout.
po-03 (Light Scribe) pattern output. WLED REST API for preset upload. ArtNet protocol specification. Hardware layout definition (user-configured).
Real-time for live installations (target 30–60 fps depending on pixel count). Preset generation is on-demand.
Converts any SVG-generating rill—Data Painter visualizations, Pattern Library tiles, cartographic layers—into optimized toolpaths for pen plotters, laser cutters, and vinyl cutters. Layer ordering, travel optimization, speed settings per material. Your watershed as a pen-plotted print. Your local pattern library laser-cut in birch plywood. The topographic contours of your backyard etched in leather.
SVG content from any generating rill. Target machine: AxiDraw, Cricut, generic plotter, laser cutter, vinyl cutter. Material settings (paper weight, wood type, acrylic thickness).
Optimized SVG with reordered paths for minimal pen-up travel. DXF export for laser cutters with layer assignments (cut, engrave, score). G-code for CNC routers. Material-specific speed/power settings. Estimated run time and material usage.
Any SVG-generating rill (po-01, fm-02, cartographic layers). Machine-specific toolpath optimization libraries. Material property databases for speed/power settings.
On-demand per export. Toolpath optimization is a one-time computation per SVG + machine configuration.
Converts terrain data from cartographic rills into CNC toolpaths or 3D-printable models. Topographic relief becomes a physical object: a mountain range on your desk, your watershed carved in walnut, the terrain around your home printed at 1:25,000 scale. Configurable vertical exaggeration, smoothing, and base thickness. The landscape, made tangible.
Terrain data from cartographic rills (c-01, c-03). Bounding box and resolution. Target: 3D print or CNC. Material and machine constraints (build volume, bit diameter, step-over).
STL/3MF mesh for 3D printing with configurable vertical exaggeration and base thickness. CNC toolpath files (G-code) with roughing and finishing passes. Estimated print time, filament usage, or machining time. Preview render of the physical object.
Cartographic terrain rills for elevation data (SRTM, USGS NED). 3D mesh generation libraries. CNC toolpath generators. Material-specific machining parameters.
On-demand per export. Terrain data is static; regenerate when bounds or resolution change.
Real-time data from any rill routed as MIDI CC values, MIDI notes, or OSC messages. Wind speed drives a synthesizer’s filter cutoff. Earthquake tremor data triggers percussive hits. Bird density modulates reverb depth. Barometric pressure bends pitch. This rill makes the data audible and performable—any DAW, any hardware synth, any OSC-capable application becomes a receiver for environmental data.
Any rill with numerical or time-series output. Mapping configuration: which data parameter drives which MIDI CC or OSC address. Range scaling and smoothing parameters.
MIDI over WebMIDI (browser-native, zero-install). OSC messages over WebSocket or UDP. Ableton Link clock sync for tempo-locked data sonification. Channel and CC assignment documentation.
Any upstream rill with numerical output. WebMIDI API (browser-native). OSC protocol specification. Ableton Link SDK for tempo synchronization.
Real-time with configurable update rate. Typical: 10–60 Hz depending on musical context. Smoothing prevents data jitter from becoming audible noise.
Assembles multi-rill outputs into publication-ready layouts: poster, broadsheet, field guide page, zine spread. Applies Forma palette and typography to create museum-quality print artifacts from data. Not a generic PDF exporter—a composition engine that understands visual hierarchy, data density, and the relationship between the content being presented. The data, made beautiful on paper.
Multiple rill outputs (maps, charts, text, images). Layout template (poster, broadsheet, field guide, zine). Forma palette and typography from fm-01 and fm-04. Print specifications (paper size, bleed, DPI).
Press-ready PDF with CMYK color, crop marks, and bleed. High-resolution PNG for digital display. InDesign-compatible markup for professional refinement. Print cost estimate based on specifications.
Any upstream rill for content. fm-01 (Palette) and fm-04 (Typography) for design language. Layout grid systems from fm-03 (Proportional System). Print specification standards (ISO 216, PDF/X).
On-demand per composition. Content updates when source rills update; regenerate to capture current data.
Generates projection-mapping-ready video content from rill data. Handles surface warping for irregular architecture, edge blending markers for multi-projector setups, and content layering for complex installations. Real-time via Syphon/NDI for live performance, or pre-rendered as HAP-encoded video for reliable playback. The building becomes a canvas; the data becomes the image projected onto it.
Visual content from poiesis rills or direct data mappings. Projection surface definition (flat, architectural scan, custom mesh). Projector specifications (resolution, throw ratio, overlap zones).
MP4/HAP-encoded video files for reliable playback. Syphon (macOS) or Spout (Windows) streams for live content. NDI streams for network-based projection. MadMapper/Resolume-compatible project files with surface mapping pre-configured.
Poiesis rills for visual content generation. Projection mapping tool APIs (MadMapper, Resolume). Syphon/Spout/NDI protocols. Surface scan data (manual or LiDAR-assisted).
Real-time for live installations. Pre-rendered content generated on-demand. Surface mapping calibration persists until physical setup changes.
Circuit board layout generation and routing from schematics. When your installation needs a custom sensor board, your LED controller needs a dedicated driver, or your environmental monitoring station needs a purpose-built data logger—PCB Designer takes a schematic and produces fabrication-ready Gerber files. KiCad scripting under the hood, with component footprint libraries and design rule checking. From concept to JLCPCB order in one rill.
Circuit schematic (KiCad format or simplified netlist). Board constraints: dimensions, layer count, component placement preferences. Target fabrication house specifications.
Gerber files (fabrication-ready). Bill of materials with component sourcing (Mouser, DigiKey, LCSC). Pick-and-place file for automated assembly. 3D board render for design verification. Design rule check report.
KiCad scripting API for layout and routing. Component footprint libraries (KiCad standard + custom). Fabrication house design rules (JLCPCB, OSH Park, PCBWay). Component distributor APIs for BOM pricing.
On-demand per design iteration. Component availability and pricing checked at export time. Design rules static per fabrication house.
Real-time status and maintenance tracking for workshop equipment: 3D printers reporting layer progress and estimated completion, CNC machines tracking spindle hours and bit wear, kilns logging firing curves against programmed schedules, laser cutters monitoring tube hours toward replacement. The workshop as a connected system—every machine reporting its state so you know what’s running, what’s finished, and what needs attention.
Machine telemetry via OctoPrint (3D printers), MQTT (generic IoT), serial/USB monitoring, or manual logging. Machine profiles with maintenance schedules and consumable lifespans.
Real-time machine status dashboard (running, idle, error, maintenance due). Job progress and estimated completion. Maintenance alerts based on hour counters and consumable tracking. Historical utilization and error logs.
OctoPrint API for 3D printer telemetry. MQTT broker for generic machine data. Kiln controller serial protocols. Manual maintenance log entries. Consumable lifespan databases (laser tubes, printer nozzles, CNC bits).
Real-time for active jobs (polling interval 5–30 seconds). Maintenance tracking updates per job completion. Consumable alerts triggered at configurable thresholds.
scaena (stage, scene, backdrop)
This is where the creation layer meets the event layer. Scaena rills design experiences that are responsive to place, time, and conditions—installations, performances, guided walks, rituals, dinners, and festivals that could only happen here and now. They also manage the technical production layer: DMX lighting protocols, spatial audio, addressable LED control, and environmental sensors that make spaces responsive. The stage is the landscape. The script is the data. The audience is anyone who shows up at the right moment.
Experience design keyed to celestial and weather conditions, site-responsive guided walks, installation logistics (power, structure, permitting), place-based ceremony frameworks, terroir-informed dinner composition, multi-day festival architecture, DMX lighting protocol management, spatial audio scene composition, addressable smart light orchestration, and environmental sensor networks for responsive installations.
Full upstream rill stack for environmental conditions (weather, celestial, phenology, ecology). DMX/ArtNet/sACN protocols for lighting (OLA, QLC+ backends). Web Audio API and Resonance Audio for spatial sound. WLED/MQTT for smart light control. TouchDesigner integration via Syphon/Spout. Environmental sensor platforms (ESP32, Raspberry Pi) via MQTT. Community event data from agora rills.
Scaena is where data becomes experience. Stack sn-01 (Experience Composer) with celestial and weather rills and a two-hour evening event has a cue sheet timed to sunset, ISS passes, and temperature drops. Layer sn-02 (Guided Walk Designer) with ecology, geology, and community knowledge rills and a trail walk changes with every season, every time of day, every weather condition. Compose sn-06 (Festival Architect) with all scaena rills and community event infrastructure and a multi-day festival unfolds with the rhythm of the site itself. Add sn-07 through sn-10 and the technical production layer—lights, audio, sensors—responds to the same data that informs the creative direction.
Stack sn-01 + sn-04 + sn-05 + sn-07 + sn-08 + celestial and weather rills for an evening that performs itself. sn-01 (Experience Composer) builds the timeline: sunset at 7:15 PM behind the ridge triggers the projection. ISS passes overhead at 7:32—the narration pauses so everyone looks up. Temperature drops below 55°F at 8:10—the gathering moves to the fire circle. sn-05 (Dinner Party Terroir) has composed a menu from what’s in season locally, with a palette drawn from autumn foliage. sn-07 (DMX Controller) fades the pathway lights through a slow amber-to-deep-blue transition as the sky darkens. sn-08 (Spatial Audio) spatializes field recordings across eight speakers, keyed to wind direction. The evening isn’t performed. It emerges from the data of the place and the moment.
| ID | Name | Description |
|---|---|---|
| sn-01 | Experience Composer | Designs timed experiences keyed to celestial events, weather, and site conditions with cue sheets |
| sn-02 | Guided Walk Designer | Site-responsive walks that change with season, time of day, weather, and ecological conditions |
| sn-03 | Installation Planner | Site-specific logistics for temporary installations: power, structure, weather, permitting, load-in timing |
| sn-04 | Ritual Designer | Place-based ceremony frameworks for solstice, equinox, harvest, and seasonal threshold gatherings |
| sn-05 | Dinner Party Terroir | Menu, palette, soundscape, and narrative arc composed from the full rill stack for a place and season |
| sn-06 | Festival Architect | Multi-day festival design combining installations, walks, meals, workshops, and ceremonies into site rhythm |
| sn-07 | DMX Controller | Lighting protocol management for installations and stages—DMX universes, fixture profiles, cue sequences |
| sn-08 | Spatial Audio | 3D audio scene composition and speaker array mapping for immersive sound environments |
| sn-09 | Smart Light Orchestrator | Addressable LED and smart lighting automation for responsive spaces—WLED, Hue, LIFX via MQTT |
| sn-10 | Sensor Canvas | Environmental sensor network for responsive installations—presence, temperature, sound level, light, motion |
Design a time-based experience informed by the full rill stack. Input: location, duration, audience, intention. Output: a structured experience score—what happens when, keyed to natural conditions. “At 7:15 PM the sun drops below the ridge. That’s when the projection begins. At 7:32 the ISS passes overhead—the narration pauses so everyone can look up.” Not a generic event template. A site-specific, time-specific, condition-specific cue sheet for an experience that could only happen here, tonight.
Location + date/time window + audience description + creative intention. Optional: specific celestial events to incorporate, weather contingencies, accessibility requirements.
Experience timeline/run sheet with cue points tied to natural events (sunset, moonrise, ISS pass, temperature threshold). Contingency plans for weather changes. Technical requirements list. Participant materials. Post-experience reflection prompts.
Celestial rills (co-05, co-06 for golden hour). Helio (t-01) for solar geometry. Weather (a-01) for conditions. po-06 (Landscape Score) for audiovisual components. Site-specific environmental data from the full rill stack.
Generated per event. Weather-dependent elements update in real time on event day. Celestial timing is pre-computed with high precision.
Creates a site-responsive guided walk keyed to what’s actually present and happening. Not a static trail guide—a walk that’s different in April than October, different at dawn than dusk, different after rain than in drought. Pulls from ecology, history, geology, sound, and community knowledge to build a stop-by-stop interpretive script with seasonal variants. The landscape narrates itself through the walker.
Trail or route (from cartographic rills or custom GPS track). Date and time of walk. Audience (family, expert naturalist, first-time visitor). Focus: ecology, geology, history, art, foraging, or mixed.
Route map with numbered stops. Stop-by-stop interpretive script with current seasonal content. Species likely to be encountered today. Historical context from archive rills. Community knowledge annotations. Estimated walk duration with pace options.
Cartographic rills for trail data. Flora and fauna rills for species presence. Birdsong rill for audio context. Archive rills for historical layers. Community knowledge (cm-07). Weather for current conditions. Phenology for seasonal timing.
Regenerated per walk date. Seasonal content shifts with phenology. Weather-dependent elements update same-day. Community annotations update as contributed.
Site-specific logistics for temporary installations: power availability and cable runs, structural attachment points and load ratings, weather exposure (wind, rain, sun), sun/shade timing throughout the day, foot traffic patterns, permitting requirements, and load-in/load-out timing optimized to light and weather. The practical backbone that turns a creative vision into something that can actually exist safely in a specific place.
Installation concept (dimensions, power requirements, structural needs). Site location. Duration. Weather tolerance thresholds. Audience size estimate.
Installation plan with site map, power distribution, structural engineering notes. Technical rider for venue/site. Weather contingency protocols. Load-in schedule optimized for light and conditions. Permitting checklist with jurisdiction-specific requirements.
Terrain (c-01) for site topography. Helio (t-01) for sun path and shade timing. Weather (a-01) and wind (a-05) for exposure. Permitting (cx-07) for regulatory requirements. Site survey data (manual or from cartographic rills).
Generated per installation. Weather monitoring updates continuously during installation period. Permitting data checked at planning time.
Templates for place-based ceremonies and seasonal rituals: solstice and equinox gatherings, harvest celebrations, planting ceremonies, water blessings, land acknowledgments. Not prescriptive—a framework of timing, orientation, and natural elements that the community fills with their own meaning. Aligns ceremony with solar geometry, lunar phase, cardinal directions, and the specific landscape where the gathering occurs.
Intention (celebration, remembrance, transition, gratitude). Location. Date or seasonal threshold. Community size. Cultural context notes.
Ritual framework with timing (keyed to solar/lunar events), spatial orientation (cardinal directions, significant landscape features), suggested natural elements (local plants, water, stone, fire), and participatory structure. Seasonal variants. Weather contingency.
Seasonal threshold data (kr-12). Helio (t-01) for solar geometry and cardinal orientations. Lunar phase (h-02). Place name etymology from archive rills. Flora rills for locally significant plants. Local cultural context from community knowledge.
Generated per ceremony. Celestial timing computed precisely for the specific date and location. Seasonal variants update with phenological calendar.
A composed experience where the meal is informed by the full rill stack. What’s in season and local from the forage rills. What ferments are possible from mt-07. What the table setting’s palette should be from the Palette Extractor. What story the Place Poem tells between courses. What music the Sonifier makes from today’s data. Not a recipe generator—a complete sensory experience design where food, color, sound, narrative, and sourcing all emerge from the same place and moment.
Location + date + guest count + dietary constraints. Intention (celebration, intimate, harvest, seasonal transition). Budget range. Kitchen capabilities.
Menu with locally sourced ingredients and seasonal justification. Table palette from fm-01. Sourcing guide (farms, markets, foraging opportunities). Playlist or soundscape from po-02. Narrative arc (place poem excerpts for between courses). Timeline from prep to service to farewell.
Seasonal harvest (fg-01) for ingredients. mt-07 (Fermentation Terroir) for beverage pairing. fm-01 (Palette) for visual design. po-04 (Place Poem) for narrative. po-02 (Sonifier) for soundscape. Local market data from agora rills.
Generated per event. Ingredient availability shifts weekly during growing season. Market sourcing verified day-of.
Designs multi-day, multi-experience festivals keyed to place and season. Combines installations, walks, meals, workshops, performances, and ceremonies into a coherent program that unfolds with the rhythm of the site. Not a conference scheduler—a festival that breathes with the landscape. Morning walks timed to birdsong peak. Afternoon workshops in the shade. Evening projections starting at sunset. The site dictates the schedule; the data informs every decision.
Site + dates + theme + capacity + available infrastructure. Creative vision and programming priorities. Budget and staffing constraints. Accessibility requirements.
Festival program with day-by-day, hour-by-hour schedule keyed to natural conditions. Site plan with installation placements, gathering zones, and flow paths. Experience timeline showing how the festival unfolds with the rhythm of the site. Technical production schedule. Logistics matrix covering power, water, waste, parking, and emergency access.
All scaena rills for component experiences. Agora rills for community event infrastructure. Full rill stack for environmental conditions, celestial timing, weather, ecology. Site survey data. Local permitting and noise ordinance databases.
Generated per festival. Weather-dependent scheduling updates daily as the event approaches. Real-time adjustments during the festival based on conditions.
Lighting protocol management for installations and stages. Manages DMX universes, fixture profiles, and cue sequences through OLA or QLC+ backends. Define your lighting rig—par cans, moving heads, LED bars, fog machines—and build cue sequences driven by rill data or manually triggered. The bridge between Slipstream’s environmental intelligence and professional stage lighting, making fixtures respond to sunset timing, weather changes, or audience presence.
Fixture manifest (DMX addresses, channel maps, fixture types). Cue list or data-driven mapping (which rill drives which lighting parameter). Universe configuration. Timing source (manual, data-driven, or show clock).
DMX data stream via OLA (Open Lighting Architecture) or QLC+ control protocol. Cue playback with crossfade timing. Real-time data-to-DMX mapping (e.g., wind speed → intensity, temperature → color temperature). Fixture status monitoring. Show file export.
OLA (Open Lighting Architecture) for DMX output. QLC+ for cue-based show control. ArtNet/sACN protocols for network DMX. Fixture profile libraries. Upstream rills for data-driven parameter mapping.
Real-time DMX output at protocol-standard 44 Hz. Cue transitions at configured crossfade rates. Data-driven mapping updates match source rill refresh rate.
3D audio scene composition and speaker array mapping for immersive sound environments. Position sound sources in three-dimensional space, define listener zones, and map outputs to physical speaker arrays. Field recordings spatialized to match their compass origin. Sonified data placed in the room where it belongs—wind sound from the east, creek sound from below, bird density overhead. Audio that occupies space the way light does.
Audio sources (field recordings, sonified rill data, music). Speaker array configuration (positions, types, amplification). Spatialization model: ambisonics, VBAP, binaural, or channel-based. Listener zone definition.
Multi-channel audio routed to physical speaker positions. Real-time spatialization parameters for Web Audio API or Resonance Audio. Ambisonic B-format for headphone rendering. Speaker calibration file. Mix visualization showing source positions and listener zones.
Web Audio API for browser-based spatial rendering. Google Resonance Audio for advanced HRTF. po-02 (Sonifier) for data-driven audio. Field recordings with geographic metadata. Speaker hardware configuration (user-defined).
Real-time audio processing at sample rate. Source position updates at rill data refresh rate. Speaker calibration is persistent per physical setup.
Addressable LED and smart lighting automation for responsive spaces. Controls WLED strips, Philips Hue, LIFX, and generic MQTT-controlled lights as a unified system. Define zones, create scenes, and bind them to rill data: hallway lights shift with circadian rhythm, studio lights match the palette extracted from today’s sky, porch lights dim when the moon is bright enough to see by. Your living space, breathing with the data of the place.
Light inventory (WLED segments, Hue bridges, LIFX devices, MQTT endpoints). Zone definitions (room, hallway, exterior). Scene configurations or data-driven bindings (which rill drives which zone’s lighting).
Unified light control across protocols: WLED REST API, Hue Bridge API, LIFX HTTP API, MQTT publish. Scene transitions with configurable fade rates. Scheduled automation tied to solar events, weather, or arbitrary rill thresholds. Energy usage tracking per zone.
WLED REST API. Philips Hue Bridge local API. LIFX HTTP API. MQTT broker for generic devices. Upstream rills for data-driven automation (helio, weather, circadian, palette). User-defined scene configurations.
Scene transitions at configured fade rate (typically 0.5–5 seconds). Data-driven updates match source rill cadence. Solar-keyed events are pre-scheduled daily.
Environmental sensor network for responsive installations. Ingests data from ESP32, Raspberry Pi, and Arduino-based sensor nodes via MQTT: presence detection, ambient temperature, sound pressure level, light intensity, motion vectors, humidity, air quality. The sensor data feeds back into the installation stack—when someone enters the room, the lights respond. When the wind picks up outside, the soundscape shifts. When temperature drops, the color palette warms. The space becomes sentient.
Sensor node telemetry via MQTT topics. Sensor types: PIR motion, ultrasonic presence, DHT22 temperature/humidity, microphone SPL, LDR light level, BME680 air quality. Node placement map.
Normalized sensor data streams available to all scaena and fabrica rills. Presence maps showing occupied zones. Environmental condition summaries. Trigger events for threshold crossings. Historical sensor data for pattern analysis.
MQTT broker (Mosquitto or cloud-hosted). ESP32/Arduino sensor firmware (ESPHome, Arduino IoT). Raspberry Pi sensor interfaces. fb-09 (PCB Designer) for custom sensor boards. Sensor calibration databases.
Real-time sensor data (50–1000 ms polling depending on sensor type). Presence detection: sub-second. Environmental conditions: 1–5 second intervals. Historical aggregation: hourly.
poiesis (making, bringing forth, creation)
Poiesis rills transform data into creative material: visual, sonic, luminous, narrative, olfactory. They take any rill’s output and render it not as a chart or a map, but as an aesthetic experience. Data becomes art-material. A year of precipitation becomes a percussive score. Bird migration density becomes a swelling chorus. Soil chemistry becomes a color palette. The watershed becomes a poem. This is the family that connects most directly to generative art practice, projection mapping, and the fundamental premise that place has an aesthetic identity waiting to be expressed.
Parameterized generative art from numerical data (p5.js, SVG, WebGL), data sonification to pitch/timbre/rhythm, addressable LED pattern generation, AI-generated poetic interpretation of assembled data, rhythmic pattern extraction from cyclical natural data, complete audiovisual landscape scores, multi-modal place identity signatures, neural style transfer from place aesthetics, AI image generation from location intelligence, and olfactory composition from botanical data.
Any upstream rill with numerical, spatial, or time-series output serves as creative input. p5.js and WebGL for visual generation. Web Audio API and Tone.js for sound. WLED/ArtNet for LED patterns. Claude API for narrative and poetic generation. Swiss Ephemeris for celestial rhythm data. Phenological databases for natural cycle timing. Phytochemistry databases for botanical aromatic profiles.
Poiesis is the aesthetic engine. Stack po-01 (Data Painter) with weather rills and earthquake data and you have a generative art installation that paints itself from environmental conditions. Layer po-02 (Sonifier) with tidal cycles and bird migration and the room fills with the sound of a place expressing its rhythms. Compose po-07 (Digital Sillage) across all available rills for a location and the result is the unique aesthetic fingerprint of that place—its “data fragrance” rendered as color, tone, glyph, and pattern. Feed any poiesis output into fabrica rills and the aesthetic becomes physical: plotted, projected, LED-mapped, printed, or performed.
Stack po-01 + po-02 + po-03 + po-05 + po-06 + environmental rills for a complete responsive art environment. po-05 (Rhythm Extractor) finds the inherent rhythms: 12.42-hour tidal cycle, dawn chorus peaking at 5:47 AM, 28-day lunar periodicity, 3-second wind gust oscillation. po-02 (Sonifier) maps tide to a bass drone that rises and falls, bird density to an upper-register chorus, wind to percussive textures. po-01 (Data Painter) renders the same data as a slowly evolving visual—color shifting with temperature, density reflecting precipitation, form tracing the tidal curve. po-03 (Light Scribe) drives 200 LEDs across the ceiling in a pattern that mirrors the bird migration heatmap. po-06 (Landscape Score) synchronizes all of it into a composition where the performer is the data and the concert hall is wherever the sensors reach.
| ID | Name | Description |
|---|---|---|
| po-01 | Data Painter | Maps numerical rill output to visual parameters: color, form, density, movement in p5.js/SVG/WebGL |
| po-02 | Sonifier | Maps rill data to sound: pitch, timbre, rhythm, spatialization via Web Audio and Tone.js |
| po-03 | Light Scribe | Maps rill data to addressable LED patterns and projection mapping content via WLED/ArtNet |
| po-04 | Place Poem | AI-generated poetic interpretation of assembled data—lyric commentary on place, not a report |
| po-05 | Rhythm Extractor | Finds inherent rhythmic patterns in natural data: tidal, circadian, seasonal, lunar periodicity |
| po-06 | Landscape Score | Composes a complete audiovisual score for a location by combining Sonifier + Data Painter + Rhythm Extractor |
| po-07 | Digital Sillage | Generates the unique “data fragrance” of a place—its signature as color + tone + glyph + pattern |
| po-08 | Style Transfer | Neural style transfer using place-derived aesthetic parameters—your photo rendered in local geology |
| po-09 | AI Image | Text-to-image generation informed by location data, forma palettes, and assembled rill intelligence |
| po-10 | Scent Composer | Olfactory composition from botanical data—the aromatic identity of a place rendered as fragrance design |
Maps any numerical rill output to visual parameters: color, form, density, movement. A parameterized generative art engine where you choose what data drives what visual dimension. Not a fixed algorithm—a mapping system. Temperature to hue. Precipitation to particle density. Wind speed to stroke velocity. Earthquake magnitude to form distortion. The visual is not a visualization of the data. It’s the data experienced as art.
Any rill with numerical output. Parameter mapping configuration: which data dimension drives which visual parameter (color, size, position, opacity, velocity, form).
p5.js sketch for interactive web display. SVG for print and plotting (fb-04). Canvas animation for screen-based installations. WebGL shader for GPU-accelerated rendering. PNG/video frame export for documentation.
Any upstream rill with numerical output. p5.js library for generative rendering. SVG specification for vector output. WebGL/Three.js for 3D. Parameter mapping is user-configured per creative intention.
Real-time for live installations (frame rate dependent on visual complexity). Batch mode for time-lapse compositions from historical data.
Maps rill data to sound parameters: pitch, timbre, rhythm, spatialization. A year of precipitation becomes a percussive score. Bird migration density becomes a swelling chorus. Earthquake magnitude becomes subharmonic rumble. Barometric pressure becomes a slowly bending drone. Not sonification-as-accessibility—sonification as aesthetic practice. The data doesn’t illustrate the sound; the data is the sound.
Any rill with time-series or spatial data. Sound mapping configuration: which data parameter drives pitch, timbre, rhythm, amplitude, spatialization. Musical constraints (scale, tempo, range).
Web Audio synthesis (browser-native, zero-install). MIDI note/CC data for external instruments via fb-06. OSC messages for DAWs and synthesis environments. Tone.js composition for structured musical output. WAV/MP3 export for offline listening.
Any upstream rill with numerical or time-series output. Web Audio API for synthesis. Tone.js for musical structure. MIDI/OSC via fb-06 for hardware routing. Musical scale and temperament databases.
Real-time for live sonification. Batch mode for composing from historical datasets. Time compression configurable (one year of data in five minutes of sound).
Maps rill data to addressable LED patterns and projection mapping content. Bird migration heatmap becomes an LED installation that pulses with species density. Solar path becomes a light sculpture that traces the sun’s arc across your room. Seismic tremor becomes a ripple of amber moving down a wall of pixels. The bridge between Slipstream’s data layer and TouchDesigner/WLED—data made luminous.
Any rill data to be rendered as light. Mapping configuration: data dimension to color, brightness, position, animation speed. Target format: LED pattern or projection content.
LED pattern data for fb-03 (LED Mapper): WLED/ArtNet-compatible. Projection content for fb-08: Syphon/NDI stream or video file. TouchDesigner CHOP export via fb-01. Pattern preview visualization.
Any upstream rill for data input. WLED color model for LED output. Syphon/NDI protocols for projection output. TouchDesigner CHOP format for bridge export. Color science libraries for perceptually uniform mapping.
Real-time for live installations (30–60 fps). Batch rendering for pre-computed projection content. Historical data playback at configurable time compression.
AI-generated poetic and narrative interpretation of a flow’s assembled data. Not a report—a lyric. “The soil remembers basalt. Sixteen species of oak breathe above it. The water table has been falling three inches a year since 1987.” Claude reads the data and writes about the place the way a naturalist-poet would: precise, vivid, grounded in observation, alive with implication. The data, made literary.
Any flow composition (assembled multi-rill data for a location). Optional: voice direction (spare, expansive, scientific, lyrical), length, focus area.
Markdown prose for screen reading. Typeset PDF via fb-07 (Print Composer). Spoken word audio via text-to-speech. Embeddable HTML fragment for integration into other compositions.
Any upstream rill data assembled into a flow. Claude API for narrative generation with system prompt optimized for naturalist-poet voice. Style calibration from exemplar texts (Lopez, Kimmerer, Leopold, Carson).
Generated on-demand. Content reflects the data state at generation time. Regenerate to capture seasonal shifts or new data.
Finds the inherent rhythmic patterns in natural data: tidal cycles as 12.42-hour pulses, circadian bird chorus timing as dawn crescendos, seasonal temperature oscillation as a slow 365-day wave, lunar periodicity as a 29.5-day heartbeat. Outputs these as musical time signatures, BPM values, and polyrhythmic patterns. The natural world already has rhythm—this rill makes it legible to musicians, composers, and installation artists.
Any rill with cyclical or time-series data: tidal, lunar, solar, phenological, weather, birdsong, seismic. Analysis window and resolution.
Detected periodicities with frequency, amplitude, and phase. Musical time signature equivalents. BPM values at multiple scales (tidal rhythm vs. annual rhythm). Polyrhythmic notation showing how different natural cycles relate. Trigger patterns for rhythmic synchronization.
Tidal data (h-01), lunar phase (h-02), solar geometry (t-01), phenological calendars, weather time series, birdsong activity patterns. FFT and autocorrelation analysis for period detection. Musical theory libraries for tempo and meter mapping.
Analysis runs on historical data windows. Real-time rhythm detection for live performance contexts. Seasonal patterns recomputed quarterly.
Composes a complete audiovisual score for a location. Combines Sonifier + Data Painter + Rhythm Extractor into a time-based composition where the performer is the data. A concert where wind drives the melody, tidal cycle sets the tempo, bird density fills the chorus, and soil moisture colors the visual field. The score is the place, performing itself across whatever time scale you choose—five minutes or five hours.
Location + time range (historical or real-time). Rill selection (which data streams to include). Aesthetic parameters: visual style, musical scale, tempo range, time compression ratio.
Synchronized audio + visual composition playable in browser. Performance score for live installation with cue marks. Video export (MP4) for screening. Separate audio (WAV) and visual (video/image sequence) tracks for independent use. Documentation of data-to-aesthetic mappings.
po-01 (Data Painter) for visual layer. po-02 (Sonifier) for audio layer. po-05 (Rhythm Extractor) for temporal structure. Multiple upstream rills for environmental data. Synchronization engine for audio-visual alignment.
Real-time for live performance. Pre-rendered for exhibition. Historical compositions generated on-demand from archived data.
Generates the unique “data fragrance” of a place—its signature across all available rills, distilled into a compact aesthetic identity. A fingerprint made of soil chemistry, sound profile, light quality, species diversity, and human history. Could manifest as a color, a tone, a glyph, a pattern—or all of them simultaneously. Two locations one mile apart will have different sillages. The same location in January and July will have different sillages. Identity through data, rendered as beauty.
Location + date (or date range for temporal sillage). Rill selection (which dimensions to include in the signature). Output modality preference: visual, sonic, multi-modal.
Multi-modal identity signature: dominant color (hex/Munsell), characteristic tone (frequency/timbre), representative glyph (generative SVG), texture pattern (procedural), and narrative essence (one sentence). Comparison view showing how this sillage differs from nearby locations or different seasons.
All available rills for the location—the more data, the richer the sillage. Dimensionality reduction (PCA/t-SNE) for distilling multi-dimensional data into aesthetic parameters. Color science, psychoacoustics, and generative geometry libraries.
Generated on-demand. Temporal sillage shifts with seasons—regenerate quarterly for seasonal identity tracking. Static sillage (geological/historical) is highly cacheable.
Neural style transfer using place-derived aesthetic parameters. Your photograph rendered in the visual language of local geology—basalt textures, ochre palette, the fractured geometry of columnar jointing. A portrait with the color palette of the autumn canopy overhead. An architectural rendering styled with the pattern library extracted from the watershed’s branching hydrology. Not generic style transfer—style transfer where the style is the place itself.
Content image (photo, rendering, illustration). Style source: location-derived (from forma and materia rills) or custom. Style intensity and blending parameters.
Styled image at configurable resolution. Style map showing which location-derived parameters influenced which image regions. Series generation (same content across multiple place-styles for comparison). Print-ready output via fb-07.
fm-01 (Palette) for color style. fm-02 (Pattern Library) for structural style. fm-05 (Material Texture) for surface style. mt-03 (Natural Pigment) for earth-tone palette. Neural style transfer models (PyTorch/TensorFlow).
On-demand per image. Style parameters from location rills are cached. Computation is GPU-intensive; typically 10–60 seconds per image depending on resolution.
Text-to-image generation informed by location data, forma palettes, and assembled rill intelligence. Not generic AI art—image generation grounded in the actual characteristics of a specific place. The prompt is enriched with soil color, dominant tree species, architectural vernacular, light quality at this latitude, and the seasonal palette of this particular week. The result is an image that could only have been generated from this place and this moment.
Creative prompt (text description of desired image). Location for place-grounding. Optional: forma palette override, specific rill data to incorporate, aspect ratio, style direction.
Generated image at configurable resolution. Prompt enrichment log showing how location data modified the base prompt. Variation series (same prompt, different location-grounding for comparison). Print-ready output via fb-07.
Image generation API (Claude vision, Stable Diffusion, or configured provider). fm-01 (Palette) for color grounding. Flora and geology rills for material/species context. po-07 (Digital Sillage) for aesthetic fingerprint. Architectural and vernacular databases for built environment context.
On-demand per generation. Location context cached from upstream rills. Generation typically 5–30 seconds per image depending on provider and resolution.
Olfactory composition from botanical data—the aromatic identity of a place rendered as fragrance design. What does this location smell like? Cedar and petrichor after rain. Wild fennel and coastal sage in the dry season. Douglas fir resin and sword fern in the understory. mt-06 (Botanical Formulation) provides the chemical profiles; Scent Composer arranges them into a fragrance architecture: top notes, heart notes, base notes, sillage. Connects directly to XO Botanicals product development.
Location + season. Botanical inventory from mt-06. Optional: fragrance family preference (woody, floral, herbal, green), intended application (personal fragrance, candle, room spray, product line).
Fragrance composition in perfumer’s notation: top/heart/base notes with proportions. Essential oil recipe with sourcing (local botanical vs. supplier). Scent story narrative connecting each note to its place origin. Seasonal variants showing how the aromatic identity shifts. Material safety data for formulation compliance.
mt-06 (Botanical Formulation) for aromatic compound profiles. Flora rills for species presence and seasonal availability. Phytochemistry databases for essential oil composition. Perfumery accord databases for complementary note pairing. IFRA safety guidelines for formulation limits.
Generated on-demand. Seasonal variants shift with phenological calendar. Botanical availability updates with growing season. Chemical profiles are static per species.
Logos — Language, Reasoning & Expression
Every rill so far produces data. Logos makes that data legible, navigable, and communicable. This is the language and reasoning layer—natural language processing, conversational interfaces, computational knowledge, knowledge graphs, trust verification, document formatting, and output presentation. Without Logos, Slipstream is a collection of sensors. With it, you can ask “How are the bees?” and receive a narrated composition drawn from six rills, formatted for your expertise level, and delivered in your preferred medium.
These rills draw from Claude API for natural language understanding and generation, Wolfram Alpha for computational knowledge, NLP libraries for text analysis, knowledge graph databases for entity relationships, cryptographic attestation for trust verification, and field-specific formatting standards from academic citation to screenplay layout to regulatory filing. Individually, a sentiment score is a number. Composed with audience calibration, document formatting, and voice rendering, it becomes a briefing that speaks to you in the right register at the right time.
lingua (tongue, language, speech)
Lingua rills are the NLP utility layer—stateless text transforms that take language in and return structured analysis. Sentiment scoring, summarization, translation, named entity recognition, tone analysis, language detection, readability metrics, and keyword extraction. These are the building blocks that every other Logos family depends on: before Logos can answer a question, Lingua must understand it. Before Stilus can format a document, Lingua must assess its readability. Pure utility, maximum composability.
Emotional valence and intensity scoring, abstractive and extractive text summarization, multilingual translation with cultural context, named entity recognition (people, places, organizations, dates, monetary amounts), tone analysis across multiple dimensions (formality, urgency, confidence, warmth), language and script identification, readability metrics (Flesch-Kincaid, Gunning Fog, SMOG), and semantic keyword extraction with topic classification.
Claude API for high-quality language understanding. Hugging Face Inference API for specialized NLP models (free tier available). spaCy for fast local NER and tokenization. VADER for rule-based sentiment as a fast fallback. Language identification libraries (fastText, langdetect). Readability formula implementations. No external data dependencies—these rills process text provided to them.
Lingua is invisible infrastructure. Stack lg-01 (Sentiment Analyzer) with community observation data and you can surface the emotional temperature of a neighborhood’s reports. Layer lg-04 (Entity Extractor) with news feeds and you automatically tag mentions of species, locations, and organizations. Compose lg-07 (Readability Scorer) with lg-03 (Audience Calibrator) and any rill’s output adjusts to a child, a general reader, a practitioner, or an expert. These rills don’t generate content—they make every other rill’s content smarter, more accessible, and more precisely targeted.
Stack lg-01 + lg-02 + lg-03 + lg-05 + lg-07 for briefings that adapt to their audience. A morning digest arrives: lg-01 scores the emotional intensity of community reports (three high-urgency water quality observations overnight). lg-02 summarizes twelve rill outputs into three paragraphs. lg-05 analyzes the tone: urgent but not alarmist, confident in the data. lg-07 scores it at grade 10 reading level. lg-03 recalibrates: for a city council member, shift to formal register and add regulatory citations; for a neighborhood parent, simplify to grade 7 and lead with health implications. Same data, same analysis, different language for different people.
| ID | Name | Description |
|---|---|---|
| lg-01 | Sentiment Analyzer | Emotional valence and intensity scoring for any text input—positive, negative, mixed, with confidence |
| lg-02 | Text Summarizer | Abstractive and extractive summarization at configurable depth—one sentence to full executive summary |
| lg-03 | Audience Calibrator | Adjusts explanation depth and register: child, general reader, practitioner, domain expert |
| lg-04 | Entity Extractor | Named entity recognition: people, places, organizations, species, dates, monetary amounts |
| lg-05 | Tone Analyzer | Multi-dimensional tone assessment: formality, urgency, confidence, warmth, persuasion |
| lg-06 | Language Detector | Identifies language and script of input text with confidence score and dialect hints |
| lg-07 | Readability Scorer | Grade level, complexity metrics (Flesch-Kincaid, Gunning Fog, SMOG), audience appropriateness |
| lg-08 | Keyword Extractor | Key terms, topic classification, semantic tagging, and auto-generated metadata from text |
Scores the emotional valence and intensity of any text input. Not just positive/negative—nuanced multi-dimensional analysis: joy, anger, sadness, fear, surprise, disgust, with intensity on a continuous scale and mixed-sentiment detection. Useful for surfacing the emotional temperature of community observations, calibrating the tone of generated narratives, and flagging high-urgency reports.
Text string (any length, any language). Optional: analysis depth (quick vs. deep), target dimensions.
Overall valence score (−1.0 to +1.0), intensity (0–1.0), discrete emotion labels with confidence, mixed-sentiment flag, and highlighted text spans driving the sentiment.
Claude API for nuanced sentiment (deep mode). VADER for fast rule-based scoring (quick mode). Hugging Face sentiment models for specialized domains.
Stateless—processes text on demand. No caching needed; each input is unique.
Abstractive and extractive summarization at configurable depth. One sentence for a notification. Three sentences for a digest entry. A full executive summary for a report. Handles single documents and multi-document synthesis—feed it twelve rill outputs and it produces a coherent narrative that captures the essential intelligence without redundancy or contradiction.
Text or array of texts. Target length (sentence count, word count, or percentage). Mode: extractive (selects key sentences) or abstractive (generates new prose).
Summary text at target length. Key points as bullet list. Coverage score (what percentage of source topics are represented). Source attribution for extractive mode.
Claude API for abstractive summarization with high coherence. Extractive algorithms (TextRank, LSA) for fast local processing. Multi-document fusion via iterative refinement.
Stateless—processes text on demand. Re-summarize when source content updates.
Adjusts explanation depth and register for four audience tiers: child (grade 3–5, concrete analogies, wonder-driven), general reader (grade 7–9, clear but not condescending), practitioner (domain vocabulary assumed, focus on implications), and domain expert (technical precision, data-forward, no hand-holding). The same rill output becomes four different texts—each true to the data, each calibrated to the reader.
Source text or rill output. Target audience tier: child, general, practitioner, expert. Optional: domain context, specific vocabulary to include or avoid.
Calibrated text at the target register. Readability score (via lg-07) confirming grade level. Vocabulary substitution log showing what was simplified or elevated.
Claude API for register transformation. lg-07 (Readability Scorer) for grade-level verification. Domain vocabulary databases for practitioner/expert tier accuracy.
Stateless transform. Re-calibrate when source content changes or audience tier shifts.
Named entity recognition tuned for Slipstream’s domains: species names (common and Latin binomial), geographic features (rivers, peaks, trails), organizations (agencies, nonprofits, tribal nations), temporal references (seasons, phenological events, dates), and monetary amounts. Feeds into knowledge graph construction (nx-04) and auto-tagging of community observations. Not generic NER—NER that knows a “chanterelle” is a species and “Willamette” is both a river and a valley.
Text string. Optional: entity types to extract (species, location, organization, temporal, monetary, all), confidence threshold.
Array of extracted entities with type, text span, character offsets, confidence score, and disambiguation (linked to canonical IDs where possible: GBIF for species, GeoNames for locations).
spaCy with custom Slipstream entity models. Claude API for complex disambiguation. GBIF species name resolver. GeoNames geographic entity database. Domain-specific gazetteers for trails, watersheds, and regulatory bodies.
Stateless transform. Entity models updated periodically as species lists and geographic databases evolve.
Multi-dimensional tone assessment beyond simple sentiment. Measures formality (casual to ceremonial), urgency (relaxed to critical), confidence (tentative to certain), warmth (clinical to intimate), and persuasion (descriptive to argumentative). Useful for calibrating generated narratives, ensuring community communications match their intended register, and detecting tonal shifts in longitudinal text data.
Text string. Optional: dimensions to analyze (all or subset), comparison baseline text.
Score per dimension (0–1.0 scale). Dominant tone label. Tonal consistency score (how uniform is the tone throughout?). Highlighted passages that drive each dimension’s score. Comparison delta if baseline provided.
Claude API for nuanced multi-dimensional analysis. Fine-tuned classifier models for rapid scoring. Linguistic feature extraction (sentence length, vocabulary complexity, hedging markers, intensifiers).
Stateless transform. No caching needed.
Identifies the language and script of input text with confidence score and dialect hints. Handles multilingual documents (paragraph-level detection), code-switching, and transliterated text. Essential for routing community observations through the correct NLP pipeline and ensuring translation rills receive properly identified source text. Supports 100+ languages with script detection (Latin, Cyrillic, CJK, Arabic, Devanagari, etc.).
Text string (minimum ~20 characters for reliable detection). Optional: granularity (document-level or paragraph-level).
Primary language (ISO 639-1 code), script (ISO 15924), confidence score, dialect/variant hint where detectable. For multilingual documents: per-paragraph language breakdown.
fastText language identification model (176 languages, offline). langdetect (Google’s language detection) as fallback. Unicode script analysis for script identification.
Stateless. Models are bundled; no external API calls needed for detection.
Grade level and complexity metrics for any text. Flesch-Kincaid grade level, Gunning Fog index, SMOG grade, Coleman-Liau index, and Automated Readability Index—five lenses on the same question: who can read this? Also computes vocabulary diversity, average sentence length, passive voice percentage, and jargon density. Feeds directly into lg-03 (Audience Calibrator) to verify that recalibrated text actually meets its target grade level.
Text string. Optional: target grade level for pass/fail assessment, jargon dictionary for domain-specific scoring.
Grade level scores (5 formulas), vocabulary diversity (type-token ratio), average sentence and word length, passive voice percentage, jargon density, and overall accessibility rating (Easy, Moderate, Difficult, Expert).
Readability formula implementations (pure computation, no API). Syllable counting via CMU Pronouncing Dictionary. Jargon dictionaries per domain (ecology, real estate, ceramics, astrology, etc.).
Stateless computation. No external dependencies. Runs entirely offline.
Extracts key terms, topic classifications, and semantic tags from text. Not just word frequency—contextual importance scoring that identifies the terms a human would use to describe what a document is about. Feeds auto-tagging of community observations, rill output indexing for search, and metadata generation for the knowledge graph (nx-04). Supports domain-aware extraction that weights ecology terms higher in a foraging observation and financial terms higher in a property report.
Text string or document. Optional: domain context for weighting, maximum keyword count, include/exclude term lists.
Ranked keyword list with relevance scores. Topic classification (broad category assignment). Semantic tag suggestions for Slipstream’s tag system. Auto-generated metadata block (title, description, category).
TF-IDF and RAKE algorithms for fast extraction. Claude API for contextual keyword identification. Domain-specific term dictionaries for relevance weighting. Slipstream tag taxonomy for tag suggestion alignment.
Stateless transform. Re-extract when source text changes. Tag taxonomy updates propagate to new extractions.
logos (word, reason, principle)
Natural language as the universal interface to the entire rill system. Instead of navigating 400+ rills by family and ID, you ask a question in plain language and Logos figures out which rills to activate, composes their outputs, and narrates the answer. “How are the bees?” activates nectar flow, inspection window, hive calendar, varroa cycle, and weather—then narrates the composite. “What if I planted blueberries here?” activates soil, zone, sun, water, and pollinator rills—then composes a scenario. The conversational front door to everything Slipstream knows.
Natural language query parsing and rill routing, voice-to-text observation intake with structured routing, contextual explanation of any rill output, hypothetical scenario composition (“What if?”), multi-location comparison across the full rill stack, educational mode with progressive depth, persistent conversation memory across sessions, and transparent multi-step reasoning with citation chains.
Claude API for natural language understanding, query decomposition, and narrative composition. Whisper API for voice transcription. The full rill catalog as a tool registry—each rill is a potential tool that Logos can invoke. Conversation memory stored in user-scoped persistent storage. Lingua rills for text analysis preprocessing.
Logos is the meta-rill family—it doesn’t produce data, it orchestrates data. Stack lo-01 (Natural Query) with the full rill catalog and you have a system where the interface is conversation rather than navigation. Layer lo-04 (What If) with scenario rills and the system becomes a planning tool. Compose lo-06 (Teach Me) with any rill output and the system becomes a teacher. lo-07 (Conversation Memory) means the system remembers what you asked last week and can reference it today. Logos transforms Slipstream from a dashboard you read into a conversation partner that knows everything the rills know.
You wake up and say: “Hey Slipstream, what’s happening?” lo-01 parses this as a general briefing request, checks your subscriptions and location, and activates: weather, golden hour, moon phase, garden soil moisture, hive inspection window, community events this week, and your daily transit horoscope. lo-07 (Conversation Memory) remembers you asked about blueberry planting last Tuesday and adds: “The soil test results you were waiting for came in—pH 5.2, ideal for blueberries.” The Narrator composes it into three paragraphs of warm, precise prose. The Voice Renderer reads it aloud while you make coffee. One question, seventeen rills, one coherent narrative—and it remembered what you care about.
| ID | Name | Description |
|---|---|---|
| lo-01 | Natural Query | Ask anything in plain language—Slipstream figures out which rills to activate and narrates the answer |
| lo-02 | Voice Memo Intake | Speak observations—transcribed, structured, and routed to the correct journal or community feed |
| lo-03 | Explain This | Point at any rill output and get a contextual, cited explanation of what it means and why it matters |
| lo-04 | What If | Hypothetical scenario composition—“What if I planted blueberries here?” becomes a multi-rill analysis |
| lo-05 | Compare Places | Side-by-side multi-rill comparison of two locations with narrative synthesis |
| lo-06 | Teach Me | Educational mode—explains the science, ecology, or methodology behind any rill’s output |
| lo-07 | Conversation Memory | Persistent context across sessions—remembers what you asked, what you care about, what changed |
| lo-08 | Reasoning Chain | Transparent multi-step reasoning with cited sources—shows the work behind every composed answer |
“How are the bees?” activates ap-01 (nectar flow), ap-02 (inspection window), ap-03 (hive calendar), ap-05 (varroa), and weather—then narrates the composite answer. The query router decomposes natural language into rill activations, executes them in parallel, assembles the results, and generates a coherent narrative response. Not keyword matching—semantic understanding of intent mapped to the rill catalog as a tool registry.
Natural language question or request. User context (location, subscriptions, history). Conversation context from lo-07.
Narrated answer composed from relevant rill outputs. Source attribution (which rills contributed). Confidence assessment. Follow-up suggestions. Raw rill data available on expansion.
Claude API for query decomposition and narrative composition. Full rill catalog as tool registry. lo-07 (Conversation Memory) for context. lo-08 (Reasoning Chain) for transparency.
Real-time query processing. Rill data freshness depends on each activated rill’s cache policy.
Speak your observation and it becomes structured data. “Saw three chanterelles under the big oak on the north trail” → transcribed via Whisper, entities extracted (species: chanterelle, quantity: 3, landmark: big oak, trail: north), geotagged, timestamped, and routed to the foraging journal with appropriate community tags. Voice-to-Feedstock: the fastest path from field observation to community intelligence.
Audio recording (voice memo). Device location for geotagging. Optional: target journal (forage, hive, general, health).
Transcribed text. Extracted entities via lg-04. Structured observation record with geolocation, timestamp, and auto-generated tags. Routing confirmation (which journal/community feed received it).
Whisper API for speech-to-text. lg-04 (Entity Extractor) for structured extraction. Device GPS for geolocation. Journal routing rules based on content classification.
Real-time transcription and processing. Observation stored immediately upon completion.
Point at any rill output and ask “why?” or “what does this mean?” The flood risk card says Zone AE. Explain This tells you: that means a 1% annual chance of flooding (the “100-year floodplain”), mandatory flood insurance if you have a federally backed mortgage, and here’s the FEMA map panel showing your parcel. Contextual, cited, calibrated to your expertise level via lg-03.
Rill output (any card, value, or visualization). User question (optional—defaults to “What does this mean?”). Audience tier from user profile.
Contextual explanation at the appropriate depth. Data citations (which source, which API, when last updated). Practical implications (“this means for you…”). Links to related rills for deeper exploration.
Claude API for explanation generation. Rill metadata (data source, methodology, last update). lg-03 (Audience Calibrator) for depth adjustment. Domain knowledge bases for accurate contextual framing.
Generated on demand. Explanation reflects the current state of the rill output being explained.
Hypothetical scenario composition. “What if I planted blueberries here?” activates soil pH, hardiness zone, sun exposure, water availability, pollinator presence, and companion planting—then composes a scenario assessment: soil pH 5.2 (ideal), Zone 8b (compatible), 6.5 hours sun (sufficient), and mason bees are present within 2 km. Not prediction—informed scenario building from assembled rill intelligence.
Natural language hypothetical question. Location context. Optional: constraints (budget, timeline, scale).
Scenario assessment with supporting evidence from activated rills. Feasibility rating. Key factors (favorable and unfavorable). Recommended next steps. Alternative scenarios if the primary is unfavorable.
Claude API for scenario decomposition and narrative synthesis. Full rill catalog as evidence sources. lo-08 (Reasoning Chain) for transparent factor analysis.
Generated on demand. Evidence freshness depends on each contributing rill’s cache policy.
“How does this compare to Portland?” Multi-rill comparison across two locations with narrative synthesis. Side-by-side: soil, climate, ecology, property values, community health, cost of living, walkability—whatever dimensions you specify or let the system choose based on what’s most different. Not just numbers in columns—a narrated comparison that explains what the differences mean for someone considering a move, an investment, or a creative project.
Two locations (current + comparison). Optional: specific dimensions to compare, purpose (moving, investing, farming, creative practice).
Side-by-side comparison table across selected dimensions. Narrative synthesis highlighting the most significant differences. Overall similarity score. Dimension-by-dimension advantage assessment.
Full rill catalog queried at both locations in parallel. Claude API for comparative narrative synthesis. lg-04 (Comparative Narrator pattern) for structured analysis.
Generated on demand. Both locations queried simultaneously for consistency.
Educational mode. Point at any rill output and say “Teach me about this.” Not just an explanation—a lesson. The flood zone card becomes a 5-minute introduction to floodplain hydrology, FEMA mapping methodology, insurance implications, and climate-driven changes. Progressive depth: start simple, go deeper on request. Links to further reading. Generates curiosity rather than just answering questions.
Any rill output as the teaching subject. Audience tier from user profile. Optional: depth preference (overview, deep dive, expert seminar).
Structured lesson: introduction, core concepts, how it works, why it matters, real-world implications, and “go deeper” paths. Interactive: ask follow-up questions within the lesson context. Quiz-yourself option for retention.
Claude API for lesson generation and interactive Q&A. Rill methodology documentation. Domain knowledge bases. lg-03 (Audience Calibrator) for depth matching. Wikipedia/educational databases for supplementary context.
Generated on demand. Lessons are session-persistent for follow-up interactions.
Persistent context across sessions. Slipstream remembers what you asked, what you cared about, what changed since your last question. “Remember I was looking into that property on Oak Street?” It remembers. “What’s changed since last Tuesday?” It knows what rills you consulted last Tuesday and surfaces what’s different now. Not conversation logging—semantic memory that understands the thread of your interests and surfaces relevant continuations unprompted.
Implicit: all conversations via lo-01 through lo-06. Explicit: “Remember this” / “Forget this” directives. Privacy: user controls what persists.
Contextual memory injection into current queries. “Since you asked” continuations. Interest thread summaries. Change detection against previously queried rill states.
User-scoped persistent storage (Supabase). Semantic embedding for memory retrieval. Change detection against cached rill states. Privacy controls per memory type.
Memory written after each conversation. Change detection runs when a remembered topic’s underlying rill data updates.
Transparent multi-step reasoning with cited sources. When Logos composes an answer from seventeen rills, the Reasoning Chain shows the work: which rills were activated and why, what each contributed, where the data came from, how confident each source is, and how the narrative was assembled. Not just “trust the answer”—a clickable audit trail from question to every data point that informed the response. Epistemic transparency as a feature, not a footnote.
Any lo-01 through lo-06 query/response pair. Optional: verbosity level (summary, detailed, full trace).
Step-by-step reasoning trace: query decomposition, rill selection rationale, per-rill data with source attribution and freshness, composition logic, confidence assessment at each step, and alternative interpretations considered.
Query processing metadata from lo-01. Per-rill provenance data (source API, last fetch timestamp, confidence). Claude reasoning trace. Composition logic documentation.
Generated alongside every composed response. Available on demand for any historical query via lo-07.
nexus (connection, bond, link)
Nexus rills find connections that individual rills can’t see. When 400+ rills each produce data about a location, the interesting question isn’t what any single rill says—it’s where the patterns converge, where correlations emerge, where causes propagate, and where risks compound. Nexus is the interpretation layer that surfaces meaning at the intersections: the relationship between soil pH and species diversity, the spatial similarity between two watersheds, the compound risk where flood zone meets infrastructure age meets demographic vulnerability.
Cross-domain statistical correlation between any two rill outputs, spatial similarity matching (find places like yours), temporal lag analysis and causal investigation with confounder identification, entity relationship graphs across species, places, people, and events, and compound risk overlay where multiple hazard factors intersect geographically.
All upstream rills as data inputs for correlation and comparison. Statistical analysis libraries (scipy, statsmodels). Graph databases (Neo4j or similar) for relationship mapping. Spatial indexing for similarity search. Risk modeling frameworks for compound overlay. No unique external data—Nexus consumes and relates data from every other family.
Nexus is where the CTI thesis becomes operational. Stack nx-01 (Cross-Domain Correlator) with any two rill families and you discover relationships invisible to either alone: bird diversity correlating with canopy age, property values tracking walkability scores, soil moisture predicting wildfire risk six months out. Layer nx-05 (Compound Risk) with flood, seismic, wildfire, and infrastructure data and the overlay reveals pockets of compounding vulnerability that no single risk assessment captures. This is Computational Tomographic Intelligence: meaning emerging at the intersections of data layers.
Stack nx-01 + nx-03 + environmental and community rills for pattern discovery. You notice that community noise complaints (cm-04) spike every March. nx-01 correlates noise reports with bird activity data and finds a 0.87 correlation: the noise complaints coincide with dawn chorus peak during spring migration. nx-03 (Causal Investigator) examines the lag: bird density increases 3–5 days before noise complaints spike, suggesting the birds aren’t causing the complaints—the early sunrise is. People sleeping with windows open for the first time hear both traffic and birdsong. The noise is there year-round; the open windows are seasonal. A hidden connection that reframes a civic complaint as a phenological signal.
| ID | Name | Description |
|---|---|---|
| nx-01 | Cross-Domain Correlator | Finds statistical relationships between any two rill outputs with significance testing |
| nx-02 | Spatial Similarity | Finds places with similar characteristics to your location across configurable dimensions |
| nx-03 | Causal Investigator | When X changed, did Y change too? Lag analysis, Granger causality, confounder identification |
| nx-04 | Network Mapper | Graph relationships between entities—species, places, people, events, concepts—as a knowledge graph |
| nx-05 | Compound Risk | Overlays multiple risk factors geographically—where do flood, fire, seismic, and infrastructure risks intersect? |
Pick any two rill outputs and discover their statistical relationship. Does bird diversity correlate with canopy age? Does soil pH predict species richness? Does walkability score track property values? The correlator runs Pearson, Spearman, and Kendall tests, checks for nonlinear relationships, and reports significance with confidence intervals. Not causation—correlation with transparent methodology, so you decide what the relationship means.
Two rill outputs with numerical or ordinal values. Geographic scope (single location, region, or national). Time window for temporal correlation.
Correlation coefficients (Pearson, Spearman, Kendall). P-values and confidence intervals. Scatter plot visualization. Nonlinearity detection. Effect size interpretation. Potential confounders flagged for nx-03 investigation.
Any two upstream rills with comparable data. Statistical libraries (scipy.stats). Spatial join engines for geographic alignment. Temporal alignment for time-series correlation.
On-demand analysis. Results cached until underlying rill data updates. Recompute when either source refreshes.
Find places with similar characteristics to yours across configurable dimensions. “Where else has this soil type, this climate, this species mix, and this community density?” Useful for: a farmer asking what other regions grow the same crops, a maker asking where else has similar clay, a mover asking which cities feel like home. Multi-dimensional similarity scoring with weighted dimensions so you control what matters most.
Reference location. Dimensions to compare (ecology, climate, demographics, economics, geology, or all). Dimension weights. Search radius or national scope.
Ranked list of similar locations with similarity scores. Per-dimension breakdown showing where they match and diverge. Map visualization of the similarity landscape. Surprise discoveries (high similarity in unexpected locations).
Full rill stack evaluated at a grid of comparison locations. Spatial indexing for efficient search. Dimensionality reduction (PCA) for multi-dimensional similarity. Pre-computed feature vectors for common dimension sets.
Pre-computed similarity indices updated monthly. On-demand analysis for custom dimension sets.
When X changed, did Y change too? And if so, was X the cause? Lag analysis to detect delayed effects (pesticide application → pollinator decline 3 weeks later). Granger causality testing to determine if one time series predicts another. Confounder identification to surface lurking variables that explain the correlation without causation. Slipstream surfaces signals, not conclusions—but this rill helps you ask the right causal questions.
Two time-series rill outputs (hypothesized cause and effect). Time window for analysis. Optional: known confounders to control for, maximum lag to test.
Lag analysis (optimal delay between cause and effect). Granger causality test results with p-values. Confounder suggestions from correlated third variables. Directed acyclic graph (DAG) of hypothesized causal structure. Confidence assessment with caveats.
Time-series data from any upstream rill pair. Granger causality implementation (statsmodels). Cross-correlation analysis. Known confounder databases by domain (ecology, economics, health).
On-demand analysis. Longer time windows improve statistical power. Re-analyze when new data extends the time series.
Graph relationships between entities—species, places, people, events, and concepts—as a navigable knowledge graph. The chanterelle connects to the Douglas fir (mycorrhizal partner), which connects to the owl (nesting habitat), which connects to the watershed (territory boundary), which connects to the salmon (spawning ground). Everything in an ecosystem is connected; this rill makes the connections visible and explorable.
Seed entity (species, place, concept, or auto-discovery from rill outputs). Relationship depth (1–3 hops). Relationship types to include (ecological, geographic, temporal, social).
Interactive force-directed graph visualization. Entity nodes with type-coded colors. Relationship edges with labels and strength. Path discovery between any two entities. Cluster identification (densely connected entity groups).
Entity extraction from lg-04. Ecological relationship databases (species interactions, food webs). Geographic ontologies. Community knowledge graph from contributed observations. Graph database (Neo4j or similar).
Graph grows continuously as new observations and rill outputs add entities and relationships. Visualization is real-time.
Overlays multiple risk factors geographically: where do flood zone, wildfire risk, seismic hazard, infrastructure age, and demographic vulnerability intersect? A location in a 100-year floodplain is one risk. That same location with aging water infrastructure, low-income population, and a hospital in the zone is a compound emergency. Individual risk rills score single hazards; Compound Risk reveals where they stack into systemic vulnerability that no single assessment captures.
Location or geographic area. Risk layers to overlay (flood, fire, seismic, infrastructure, demographic, environmental, or all). Weighting scheme (equal or custom).
Compound risk score (1–10) with per-layer contribution. Risk overlay map showing geographic intersection zones. Hotspot identification (highest compound risk areas). Scenario modeling: how does compound risk change if one layer shifts?
Risk-producing rills: vl-02 (flood), vl-05 (insurance risk), seismic data, wildfire models. Infrastructure data from grid rills. Demographic vulnerability from census and community health rills. Spatial overlay engines for geographic intersection.
Risk layers refresh at their individual cadences. Compound overlay recomputes when any contributing layer updates. Scenario models are on-demand.
wolfram (wolf-raven, Germanic)
The universal computation rill. Ask any mathematical, scientific, or data question in natural language and get a precise, computed answer. Unit conversions, statistical analysis, chemical properties, astronomical calculations, geodetic computations, number theory, physical constants—everything that has a computable answer. Wolfram Alpha under the hood, with Slipstream wrapping the raw computational power in context-aware interfaces that connect results back to the rill ecosystem.
Natural language math and science queries, context-aware unit conversions, statistical significance testing and distribution analysis, chemical compound data and reaction balancing, precise orbital mechanics and eclipse computation, geodetic distance/bearing/area with ellipsoidal corrections, prime factorization and sequence exploration, periodic table and physical constants, structured knowledge queries, and symbolic equation solving with step-by-step derivation.
Wolfram Alpha Full Results API (paid, comprehensive). Wolfram Alpha Short Answers API (free tier for quick queries). Wolfram Language/Cloud for programmable computation. Supplementary: scipy/numpy for local statistical computation, periodic table databases, NIST physical constants, geodetic libraries (GeoPy, pyproj).
Wolfram is the rill that answers questions no other rill was designed for. Stack w-05 (Astronomical Calculator) with astra rills and you get ecliptic coordinates computed to arc-second precision alongside the interpretive horoscope. Layer w-06 (Geodetic Computer) with cartographic rills and distance calculations account for Earth’s actual ellipsoidal shape. Compose w-03 (Statistical Analysis) with any data-producing rill and you can test whether the pattern you see is significant or noise. Wolfram fills the computational gaps between specialized rills with a universal solver.
Stack w-01 + w-06 + w-05 + cartographic and celestial rills for precision at every layer. You’re planning an eclipse viewing location. w-05 computes the eclipse path with arc-second precision: center line crosses your county at 10:47:23 AM PDT, duration of totality 2 minutes 18 seconds at the optimal viewing point. w-06 calculates the geodetic distance from your home to that point: 47.3 km bearing 247° (accounting for Earth’s ellipsoidal shape, not the flat-map approximation). w-01 answers the follow-up: “What’s the altitude of the sun at totality?” — 58.4° above the horizon, so the ridgeline to the south won’t obstruct the view. Every number computed, not estimated.
| ID | Name | Description |
|---|---|---|
| w-01 | Compute Anything | Natural language math/science/data query via Wolfram Alpha—the universal computational interface |
| w-02 | Unit Converter | Context-aware conversions: acres to hectares, Fahrenheit to Celsius, board-feet to cubic meters |
| w-03 | Statistical Analysis | Significance testing, distributions, confidence intervals, and correlation analysis on rill data |
| w-04 | Chemical Properties | Compound data, reaction balancing, molecular visualization, and safety information |
| w-05 | Astronomical Calculator | Precise orbital mechanics, eclipse paths, conjunction timing, and celestial coordinate transforms |
| w-06 | Geodetic Computer | Distance, bearing, and area calculations with ellipsoidal corrections on the WGS84 geoid |
| w-07 | Number Theory | Prime factorization, integer sequences (OEIS), mathematical explorations, and pattern discovery |
| w-08 | Science Reference | Periodic table, physical constants (NIST), material properties, and element/isotope data |
| w-09 | Data Query | Structured queries against curated knowledge databases—countries, cities, demographics, economics |
| w-10 | Formula Solver | Symbolic equation solving with step-by-step derivation—algebra, calculus, differential equations |
The universal computational interface. Ask any mathematical, scientific, or factual question in natural language and get a precise, computed answer from Wolfram Alpha. “What’s the population density within 50 km of Portland?” “How many joules in a kilowatt-hour?” “What’s the orbital period of Jupiter?” The answer isn’t retrieved from a database—it’s computed from first principles and curated knowledge.
Natural language query. Optional: output format preference (numerical, textual, visual), unit system (metric, imperial).
Computed answer with units. Step-by-step solution where applicable. Visualizations (plots, diagrams, structures). Related computations and alternative interpretations. Source attribution and methodology.
Wolfram Alpha Full Results API (primary). Wolfram Alpha Short Answers API for quick queries. Wolfram Language/Cloud for programmable computation.
Computed on demand. Mathematical results are exact and don’t expire. Factual queries reflect Wolfram’s curated data (updated continuously).
Context-aware unit conversions with Wolfram precision. Not just “miles to kilometers”—conversions that understand domain context. A ceramicist asks about cone temperatures and gets Fahrenheit, Celsius, and Orton cone equivalents. A forester asks about board-feet and gets cubic meters with lumber grade adjustments. A farmer asks about acre-feet of water and gets liters, gallons, and irrigation hours at their flow rate.
Value with source units. Target units (or “all common”). Optional: domain context for specialized conversions.
Converted value(s) with full precision. Domain-specific equivalents where applicable. Conversion factor and formula. Quick-reference for the most common conversions in the detected domain.
Wolfram Alpha for comprehensive unit handling. Domain-specific conversion tables (ceramics: cone charts, forestry: board-foot calculators, agriculture: irrigation equivalents).
Exact computation. No expiration. Currency conversions use daily exchange rates.
Is the pattern real or noise? Feed any rill’s data into significance testing: t-tests, chi-squared, ANOVA, Mann-Whitney U. Fit distributions to observed data. Compute confidence intervals. Run regression analysis. When nx-01 finds a correlation, w-03 tests whether it’s statistically significant. Slipstream surfaces signals—this rill tells you which ones to trust.
Numerical dataset(s) from any rill. Analysis type: descriptive statistics, hypothesis test, distribution fitting, regression, or auto-detect. Significance level (default: 0.05).
Test results with p-values, effect sizes, and confidence intervals. Distribution fit with goodness-of-fit metrics. Regression coefficients with R². Visualization: histogram, Q-Q plot, residual plot. Plain-language interpretation.
scipy.stats for comprehensive statistical testing. numpy for numerical computation. Wolfram Alpha for verification of complex analyses. lg-03 for plain-language interpretation at the user’s level.
On-demand analysis. Recompute when underlying dataset updates.
Compound data, reaction balancing, molecular visualization, and safety information. What’s the molecular weight of linalool? Balance the combustion equation for methane. Show the structure of caffeine. What’s the LD50 of permethrin? Connects to mt-02 (Glaze Chemistry) for ceramic oxide calculations and mt-06 (Botanical Formulation) for essential oil compound profiles.
Compound name, formula, or CAS number. Query type: properties, reaction balancing, structure visualization, safety data.
Physical and chemical properties (MW, melting/boiling point, density, solubility). Balanced reaction equations. 2D/3D molecular structure visualization. Safety data (GHS classification, LD50, exposure limits). Cross-references to PubChem and ChemSpider.
Wolfram Alpha chemical data. PubChem database (free, NIH). Safety data sheets (GHS-compliant). Ceramic oxide chemistry tables for UMF calculations.
Chemical properties are static. Safety data updates with regulatory changes. PubChem data refreshes continuously.
Precise orbital mechanics, eclipse paths, conjunction timing, and celestial coordinate transforms. When the cosmo rills provide observational astronomy and the astra rills provide interpretive astrology, Wolfram’s astronomical calculator provides the mathematical backbone: ephemeris computation to arc-second precision, eclipse path geometry, planetary conjunction timing, and coordinate system transforms (equatorial, ecliptic, horizontal, galactic).
Astronomical query: body positions, event timing, coordinate transforms. Location and time for observer-specific calculations.
Computed positions with arc-second precision. Event timing (eclipses, conjunctions, occultations) with location-specific circumstances. Coordinate transforms between reference frames. Visibility windows and altitude/azimuth plots.
Wolfram Alpha astronomical engine. JPL Horizons system for high-precision ephemerides. Swiss Ephemeris for astrological cross-reference. IAU/SOFA libraries for coordinate transforms.
Computed on demand. Orbital mechanics are deterministic—results are exact for any requested time.
Distance, bearing, and area calculations with ellipsoidal corrections on the WGS84 geoid. Not the flat-map approximation—the real distance accounting for Earth’s oblate shape. Vincenty and Karney algorithms for sub-meter accuracy over any distance. Polygon area computation using the geodesic method. Bearing and azimuth with convergence correction. The math underneath every map-based rill, exposed as a direct tool.
Two or more coordinate pairs (lat/lon). Calculation type: distance, bearing, area, midpoint, destination point from bearing+distance.
Geodesic distance with sub-meter precision. Forward and reverse azimuths. Polygon area in m², hectares, and acres. Midpoint coordinates. Destination point from start+bearing+distance. Comparison to flat-map approximation showing the error.
GeographicLib (Karney inverse/direct algorithms). pyproj for coordinate reference system transforms. WGS84 ellipsoid parameters. Wolfram Alpha for verification.
Pure computation. No external data dependencies. Results are mathematically exact.
Prime factorization, integer sequences, and mathematical explorations. Is this number prime? What’s the 1000th Fibonacci number? Find the sequence in OEIS. Compute the GCD of these measurements. Generate the Collatz sequence from this starting point. The pure mathematics rill—for when the question is about numbers themselves rather than what they measure. Surprisingly useful for pattern recognition in natural data where cycles and ratios appear.
Integer(s) or mathematical expression. Query type: factorization, primality test, sequence identification, GCD/LCM, modular arithmetic.
Complete prime factorization. Primality certificate. OEIS sequence identification with formula. GCD/LCM computation. Sequence visualization. Related mathematical properties and theorems.
Wolfram Alpha number theory engine. OEIS (Online Encyclopedia of Integer Sequences). SymPy for local symbolic computation. Primality testing algorithms (Miller-Rabin, AKS).
Pure computation. OEIS database updates as new sequences are catalogued.
Periodic table, physical constants, material properties, and element/isotope data. What’s the electron configuration of iron? What’s the speed of sound in water at 20°C? What’s the thermal conductivity of Douglas fir? NIST-sourced physical constants with full uncertainty. Material property databases for engineering and craft applications. The reference shelf of a scientist’s lab, always current and always precise.
Element, compound, or material name. Property query: atomic data, physical constants, material properties. Optional: conditions (temperature, pressure).
Requested property with units, uncertainty, and conditions. Element: full atomic data (mass, configuration, isotopes, abundance). Constants: CODATA recommended values. Materials: thermal, mechanical, optical, electrical properties. Source citation.
NIST physical constants (CODATA, free). Periodic table databases. Material property databases (MatWeb, engineering handbooks). Wolfram Alpha for cross-verification.
Physical constants update with CODATA releases (every ~4 years). Material properties are static per material. Element data is definitive.
Structured queries against Wolfram’s curated knowledge databases. “What’s the GDP of Oregon?” “List countries with similar population density to the Netherlands.” “What’s the elevation of every state capital?” Unlike a search engine, the answers are computed from structured data—sortable, filterable, and comparable. The structured knowledge layer that fills gaps when specialized rills don’t cover a domain.
Natural language data query. Optional: output format (table, chart, single value), comparison set, filtering criteria.
Structured data response: tables, ranked lists, single values with context. Comparison visualizations. Data source attribution. Download as CSV for further analysis.
Wolfram Knowledgebase (curated, continuously updated). World Bank Open Data. UN Statistics Division. CIA World Factbook. National statistical agencies.
Wolfram’s curated data updates continuously. Economic data typically lags 1–2 quarters. Demographic data lags 1–2 years.
Symbolic equation solving with step-by-step derivation. Solve for x. Integrate this function. Find the eigenvalues. Compute the Laplace transform. Not just the answer—the work, shown clearly enough to learn from. Connects to lo-06 (Teach Me) for educational contexts where understanding the derivation matters as much as the result. Algebra through differential equations, with every step explained.
Mathematical expression or equation. Operation: solve, simplify, differentiate, integrate, factor, expand, series expansion. Variables to solve for.
Solution(s) with domain restrictions. Step-by-step derivation showing each algebraic manipulation. Visualization (function plot, solution set). Alternative solution methods when applicable. LaTeX-formatted output for publication.
Wolfram Alpha symbolic engine. SymPy for local computation. Maxima CAS as fallback. Step-by-step derivation via Wolfram|Alpha Pro or equivalent.
Pure computation. Results are mathematically exact. No external data dependencies.
fides (trust, faith, confidence)
How do you trust data contributed by strangers? Fides is the trust and governance infrastructure for community intelligence. When someone posts “I saw chanterelles at the trailhead,” fides determines: is this person credible (reputation score from verification history), can the claim be cross-verified (geological and ecological plausibility), where did it come from (provenance chain), and how should the community govern access to this kind of information (governance framework). Trust isn’t binary—it’s a spectrum, and fides makes it transparent.
Contributor reputation scoring based on verification history, multi-source cross-checking pipelines, data lineage and provenance tracking, community governance tools (proposals, voting, moderation), collective data ownership models with contribution/benefit tracking, and cryptographic attestation for observation time, location, and content integrity.
Community contribution history (Supabase). Cross-verification against authoritative rill data (e.g., geological plausibility from geology rills, ecological plausibility from flora rills). Cryptographic signing via Web Crypto API. Governance state stored in community-scoped databases. No external trust APIs—fides builds trust from internal verification patterns.
Without fides, community intelligence is just crowdsourced noise. Stack fd-01 (Reputation System) with communis rills and every community observation carries a trust score. Layer fd-02 (Verification Pipeline) with geology and ecology rills and the chanterelle sighting gets cross-checked: is there suitable mycorrhizal host species nearby? Does the soil type support chanterelle growth? Is this the right season? fd-03 (Provenance Tracker) creates an audit trail: this claim was made by a user with a 0.92 reputation score, cross-verified against three independent data sources, with GPS attestation. Trust becomes a composable data type rather than an assumption.
A community member reports: “Water smells off at the creek behind the middle school.” fd-01 checks reputation: contributor score 0.78, 23 previous observations, 19 verified. fd-06 (Attestation) confirms GPS coordinates match the creek location with timestamp integrity. fd-02 cross-checks: EPA water quality data shows the upstream monitoring station recorded elevated turbidity 48 hours ago; three other community members have reported similar observations within 2 km in the last week. fd-03 logs the provenance chain: community observation → GPS attestation → EPA cross-reference → community corroboration. Result: confidence 0.89, flagged for the community water monitoring dashboard. Not one person’s opinion—a verified, cross-referenced, provenanced signal.
| ID | Name | Description |
|---|---|---|
| fd-01 | Reputation System | Contributor trust scores based on verification history, accuracy, and community standing |
| fd-02 | Verification Pipeline | Multi-source cross-checking for community observations against authoritative rill data |
| fd-03 | Provenance Tracker | Data lineage—where did this claim come from? How was it verified? Who contributed? |
| fd-04 | Governance Framework | Community decision-making tools: proposals, voting, moderation, access policies |
| fd-05 | Data Cooperative | Collective data ownership models with contribution tracking and benefit distribution |
| fd-06 | Attestation Engine | Cryptographic proof of observation time, location, and content integrity |
Contributor trust scores built from verification history. Each community observation that gets cross-verified increases the contributor’s reputation. Observations that contradict authoritative data decrease it. Domain-specific: a contributor with high foraging accuracy might have low reputation for water quality reports if they haven’t contributed there. Reputation is earned through consistent, verifiable contributions—not social popularity.
Contributor ID. Domain (foraging, water, trails, noise, general). Verification history from fd-02.
Overall reputation score (0–1.0). Per-domain scores. Verification success rate. Contribution count and consistency. Reputation trajectory (improving, stable, declining).
Community contribution database (Supabase). fd-02 verification results. Contribution frequency and recency. Domain-specific accuracy tracking.
Updated after each verification event. Score recomputed incrementally, not from scratch.
Multi-source cross-checking for community observations. When someone reports chanterelles at a location, the pipeline checks: is the soil type appropriate (SSURGO)? Are mycorrhizal host trees present (flora rills)? Is this the right season (phenology)? Have other observers reported similar sightings nearby? The result is a confidence score based on how many independent sources corroborate the claim—not whether we “believe” the person, but whether the data is consistent with what we know.
Community observation (claim, location, timestamp, contributor). Verification sources to check against (auto-selected by claim type).
Confidence score (0–1.0). Per-source verification results. Corroboration count. Contradiction flags. Overall assessment: Verified, Plausible, Unverified, Contradicted.
Authoritative rill data for cross-reference (geology, flora, weather, phenology). Community observation database for corroboration. fd-06 for attestation verification.
Runs automatically on each new community observation. Re-verifies when new corroborating data arrives.
Data lineage for every claim in the system. Where did this data come from? A federal API? A community contributor? A computed inference? How was it verified? By cross-reference, by corroboration, by attestation? When was it last updated? Provenance Tracker makes the entire data supply chain transparent—clickable from any data point back to its ultimate source. Epistemic hygiene as infrastructure.
Any data point, observation, or composed result in the system. Query: “Where did this come from?”
Provenance chain: original source → transformations → verifications → current presentation. Timestamps at each step. Contributor attribution. Data freshness. Confidence assessment at each link in the chain.
Rill metadata (source API, fetch timestamp, cache status). Community contribution records. fd-02 verification logs. fd-06 attestation records. Composition logs from lo-08.
Provenance recorded at data creation and each transformation. Chain is append-only and immutable.
Community decision-making tools for the data commons. Proposals (“Should we make foraging locations fuzzy within 500m for conservation?”), voting (reputation-weighted or one-person-one-vote), moderation (flagging, appeals, conflict resolution), and access policies (who can see what, under what conditions). Not imposed governance—a framework the community configures for itself, with transparent rules and auditable decisions.
Governance action: proposal submission, vote casting, moderation flag, policy change request. Community context (who has standing, what rules apply).
Proposal status and vote tallies. Policy documents with version history. Moderation decisions with rationale. Access control matrices. Governance health metrics (participation rate, decision velocity).
Community governance state (Supabase). fd-01 reputation for weighted voting. Proposal and voting history. Moderation logs. Policy version control.
Real-time for active proposals. Policy changes effective immediately upon passing threshold. Governance metrics aggregated daily.
Collective data ownership models. When a community contributes observations that collectively create valuable intelligence (a neighborhood air quality map, a regional foraging calendar, a community noise profile), who owns that composite? Data Cooperative tracks contributions, models shared ownership, and distributes benefits. If the composite data is licensed commercially, contributors share proportionally. Data dignity as infrastructure, not an afterthought.
Community data corpus. Contribution records. Licensing/usage requests. Benefit distribution model (equal, proportional, threshold-based).
Contribution ledger (who contributed what, when). Ownership shares per contributor. Usage/licensing agreements. Benefit distribution records. Community data value metrics.
Community contribution database. fd-03 provenance records for attribution. Licensing framework templates. Benefit distribution engine. Smart contract patterns for automated distribution.
Contribution ledger updates with each new observation. Ownership shares recomputed periodically. Benefit distribution triggered by licensing events.
Cryptographic proof of observation time, location, and content integrity. When a community member submits an observation, the Attestation Engine signs it with a timestamp, GPS coordinates, and content hash. The attestation is tamper-evident: if the observation is modified after submission, the signature invalidates. Not blockchain—standard cryptographic signing (Web Crypto API) with a lightweight verification chain. Trust through math, not authority.
Observation content (text, photo, audio). Device GPS coordinates. Device timestamp. Contributor identity (public key).
Signed attestation: content hash + GPS + timestamp + contributor signature. Verification endpoint for third-party validation. Tamper-evidence flag on any subsequent modification. Attestation chain linking related observations.
Web Crypto API for signing (SubtleCrypto, ECDSA P-256). GPS from device. NTP-synchronized timestamps. Content hashing (SHA-256). Verification stored in Supabase with signature integrity checks.
Attestation created at observation time. Immutable once signed. Verification is instantaneous (signature check is O(1)).
stilus (pen, stylus, writing instrument)
Every field has formatting conventions that gatekeep access. Academic papers require APA or Chicago style. Legal briefs demand Bluebook citation. Screenplays follow industry-standard formatting that signals professionalism before a word is read. Grant proposals fail on formatting violations before the science is evaluated. Stilus democratizes this knowledge: ten rills covering every major document convention from academic to regulatory, ensuring that brilliant content isn’t rejected for incorrect margins.
Academic citation styles (APA, MLA, Chicago, IEEE, Vancouver, and 10+ more), legal document standards (Bluebook, court-specific formatting), screenplay and teleplay formatting (feature, TV, stage), book manuscript conventions (Shunn format, query letters), grant proposal requirements (NIH, NSF, NEA, NEH), business documents (pitch decks, SOWs, MOUs), technical writing standards (IMRaD, Diátaxis, RFC), journalistic style (AP Style, press releases), correspondence protocol, and regulatory filing formats (FDA, SEC, USPTO).
Style guide databases (APA 7th, Chicago 17th, MLA 9th, Bluebook 21st). Format specification documents per field. Agency-specific formatting requirements (NIH, NSF, NEA, FDA, SEC, USPTO). Industry standard templates (Final Draft for screenplays, Shunn for manuscripts). Claude API for intelligent formatting application and validation.
Stilus makes content portable across professional contexts. Stack st-01 (Academic Formats) with research rills and a community ecology study gets properly formatted for peer review submission. Layer st-05 (Grant & Proposal) with project planning rills and a restoration project becomes an NEA-compliant funding application. Compose st-06 (Business & Professional) with market analysis rills and the investor update for XO Botanicals arrives in the format VCs expect. The content is the same; stilus ensures it speaks the language of its audience.
Stack st-05 + research rills + project planning rills for a funding-ready proposal. You’re applying for an NEA grant to fund a data-driven public art installation. st-05 knows the NEA format: project narrative (3 pages max), budget justification, timeline, work samples, organizational history. It validates margins (1 inch), font (12pt), line spacing (single for budget, double for narrative). The content comes from your installation planning rills; st-05 ensures the formatting won’t trigger a desk rejection. The grant panel reads your vision. They never see the formatting—which is exactly the point.
| ID | Name | Description |
|---|---|---|
| st-01 | Academic Formats | APA, MLA, Chicago, IEEE, Vancouver + 10 more citation styles with automatic formatting |
| st-02 | Legal Document Standards | Bluebook citation, court-specific brief formatting, contract drafting conventions |
| st-03 | Screenplay & Teleplay | Feature, TV, stage play, and audio drama formatting with page-time analysis |
| st-04 | Book Manuscript Standards | Query letters, synopses, Shunn manuscript format, book proposals |
| st-05 | Grant & Proposal Format | NIH, NSF, NEA, NEH—agency-specific formatting validation and template generation |
| st-06 | Business & Professional | Pitch decks, SOWs, MOUs, investor updates, board decks with industry conventions |
| st-07 | Technical & Scientific | IMRaD validation, Diátaxis documentation framework, RFC format, API documentation |
| st-08 | Journalistic Standards | AP Style enforcement, inverted pyramid structure, press release formatting |
| st-09 | Correspondence | Business letters, diplomatic protocol, international conventions, formal communication |
| st-10 | Regulatory & Compliance | FDA submissions, SEC filings, USPTO patent claims, EPA environmental impact formatting |
APA 7th, MLA 9th, Chicago 17th, IEEE, Vancouver, AMA, Turabian, Harvard, CSE, OSCOLA, AGLC, Bluebook (academic), AAA, ASA, NLM—fifteen citation styles with automatic bibliography generation, in-text citation formatting, and compliance validation. Input your references and prose; output arrives formatted for submission. Detects and flags common errors: hanging indent violations, italics inconsistencies, DOI formatting, and missing fields.
Document text with reference markers. Bibliography entries (BibTeX, RIS, or manual). Target style.
Formatted document with in-text citations and bibliography. Compliance report. Error flags with fix suggestions.
Style guide rule databases. CrossRef DOI resolution. Claude API for intelligent formatting decisions in ambiguous cases.
Style rules update with new guide editions. DOI resolution is real-time.
Bluebook citation formatting for law review articles and court filings. Court-specific brief formatting (federal, state, appellate—each with different margin, font, and page limit rules). Contract drafting conventions with defined terms, recitals, and boilerplate. Not legal advice—formatting intelligence that ensures the document looks right before a lawyer reviews the substance.
Legal document text. Document type (brief, motion, contract, memo). Jurisdiction for court-specific rules.
Formatted document meeting jurisdiction-specific rules. Citation compliance check. Page count and word count verification against court limits.
Bluebook 21st Edition rules. Court-specific local rules databases. Contract drafting conventions. Claude API for context-sensitive formatting.
Court rules update periodically. Bluebook updates with new editions.
Industry-standard screenplay formatting: scene headings (INT./EXT.), action lines, character names, dialogue, parentheticals, transitions. Supports feature film, television (with act breaks), stage play (Samuel French format), and audio drama. Page-time analysis: approximately one minute per properly formatted page. Dual dialogue, montage sequences, and intercut formatting. The invisible formatting that signals “this writer is professional” before a single word of dialogue is read.
Screenplay text (plain text or Fountain markup). Format type: feature, TV pilot, TV episode, stage play, audio drama.
Industry-formatted PDF (Courier 12pt, specific margins). Page count with runtime estimate. Scene breakdown (INT/EXT count, location list). Character dialogue distribution analysis.
Industry formatting standards. Fountain markup parser. Final Draft-compatible export. Page-time calculation models.
Formatting standards are stable. Regenerate on content changes.
Shunn manuscript format for fiction submissions (Times New Roman 12pt, double-spaced, 1-inch margins, header with name/title/page). Query letter templates with personalization guidance. Synopsis formatting at multiple lengths (1-page, 3-page, full). Book proposal structure for nonfiction (overview, market analysis, chapter summaries, author platform). The gatekeeping conventions that determine whether an agent reads past page one.
Manuscript text. Document type: short story, novel manuscript, query letter, synopsis, book proposal. Target market (genre, publisher type).
Formatted manuscript meeting Shunn or publisher-specific standards. Word count and estimated page count. Query letter draft with personalization placeholders. Synopsis at requested length.
Shunn manuscript format specification. Publisher submission guidelines databases. Query letter best practices. Claude API for synopsis compression.
Format standards are stable. Publisher guidelines checked at submission time.
Agency-specific formatting validation for grant applications. NIH (R01 page limits, biosketch format, specific aims structure), NSF (project summary/description, broader impacts, data management plan), NEA (project narrative, work samples, budget justification), NEH (narrative, bibliography, appendices). Each agency has different rules; this rill knows them all. Validates before submission so formatting violations don’t trigger desk rejection.
Proposal content. Target agency and program. Budget figures. Supporting documents list.
Formatted proposal meeting agency requirements. Compliance checklist with pass/fail per requirement. Page count verification. Budget template in agency format. Missing element alerts.
Agency-specific formatting guides (NIH, NSF, NEA, NEH, USDA, DOE, etc.). Grants.gov submission requirements. Claude API for intelligent section organization.
Agency guidelines update annually. Re-validate when guidelines change or new funding opportunities post.
Pitch decks (10–15 slide structure, story arc, data visualization standards), statements of work (scope, deliverables, timeline, payment terms), MOUs (party identification, obligations, term, termination), investor updates (KPIs, burn rate, milestones, ask), and board decks (executive summary, financials, strategic decisions). Each format has conventions that signal competence to its audience. This rill knows them.
Business content. Document type: pitch deck, SOW, MOU, investor update, board deck, executive summary. Industry context.
Formatted document in the expected structure. Slide deck with proper flow. Financial tables in standard presentation. Action items and decision points highlighted.
Business document conventions by industry. Pitch deck best practices. Legal template libraries for SOW/MOU. Claude API for structure optimization.
Business conventions evolve slowly. Templates update with industry trends.
IMRaD structure validation for scientific papers (Introduction, Methods, Results, and Discussion). Diátaxis documentation framework (tutorials, how-to guides, reference, explanation). RFC format for technical standards. API documentation conventions (OpenAPI/Swagger). README templates. Technical writing that serves its reader rather than its author.
Technical document. Format type: scientific paper, documentation, RFC, API docs, README. Target audience.
Structured document meeting format conventions. Section completeness check. Terminology consistency validation. Cross-reference integrity. Reading order optimization.
IMRaD structural rules. Diátaxis framework specification. IETF RFC formatting guidelines. OpenAPI specification. Claude API for structural analysis.
Technical standards update with new specification versions. Re-validate on major format updates.
AP Style enforcement for news writing: capitalization, abbreviation, number formatting, dateline conventions. Inverted pyramid structure validation (most important information first, supporting details in descending order of importance). Press release formatting (FOR IMMEDIATE RELEASE, dateline, boilerplate, contact information). Op-ed structure. Feature article conventions. The formatting that makes writing look like journalism rather than a blog post.
Article text. Type: news story, press release, op-ed, feature, brief. Publication context.
AP Style-compliant text with corrections highlighted. Structure analysis (inverted pyramid compliance). Press release in standard format. Word count and estimated column inches.
AP Stylebook rules database. Press release format standards. Inverted pyramid structural analysis. Claude API for style correction suggestions.
AP Stylebook updates annually. Format conventions are stable.
Business letter formatting (block, modified block, semi-block), diplomatic protocol (forms of address for officials, ambassadors, heads of state), international correspondence conventions (date formatting, salutation norms by culture), and formal communication structure. Knows that a letter to a Japanese business partner has different salutation conventions than a letter to a French minister. The invisible protocol that prevents unintended offense.
Letter content. Recipient role/title. Format type: business letter, diplomatic note, formal invitation, thank-you. Cultural context.
Formatted letter with correct salutation, closing, and structural conventions. Forms of address lookup. Cultural guidance notes. Date and number formatting for the recipient’s locale.
Diplomatic protocol guides. Business correspondence conventions by culture. Forms of address databases. International date/number formatting standards (ISO 8601, locale-specific).
Protocol conventions are stable. Forms of address update with personnel changes.
FDA drug and device submissions (510(k), NDA, IND formatting), SEC filings (10-K, 10-Q, 8-K, S-1 structure), USPTO patent claims (independent and dependent claim formatting, specification structure), and EPA environmental impact statement formatting. The most consequential formatting in the system—a misformatted FDA submission can delay a drug by months. Regulatory formatting intelligence that ensures the paperwork doesn’t become the bottleneck.
Regulatory document content. Filing type and target agency. Jurisdiction. Supporting data references.
Formatted submission meeting agency-specific requirements. Section completeness validation. Cross-reference integrity check. Submission checklist with all required attachments.
Agency formatting guides (FDA eCTD, SEC EDGAR, USPTO MPEP, EPA NEPA). Regulatory submission templates. Claude API for intelligent section organization and cross-reference validation.
Regulatory formatting rules update with agency guidance revisions. Re-validate when new guidance is published.
formato (format / shape given)
Every rill in Slipstream produces data; formato is where that data becomes something a human can hold—a printed field guide, a narrated audio digest, a gallery-quality map, a bento-box dashboard. Where stilus enforces how text is formatted, formato decides what form the output takes: prose, chart, PDF, projection, voice. It is the final mile between intelligence and experience.
aer-03 → terr-02 → st-07 → fo-03 → fo-05
Air quality forecast feeds terrain context, stilus enforces technical formatting, report builder assembles the document, field guide generator produces a printable pocket edition.
| ID | Name | One-liner |
|---|---|---|
| fo-01 | Narrator | AI prose generation at three depths: Chroma, Lux, Nocturne |
| fo-02 | Dataviz Engine | Shared visualization library—gauges, sparklines, radar, heatmaps |
| fo-03 | Report Builder | PDF / DOCX / HTML document generation from composed rill outputs |
| fo-04 | Dashboard Composer | Bento-box UI with drag-and-drop rill arrangement and saved layouts |
| fo-05 | Field Guide Generator | Printable, location-specific, seasonal pocket guides |
| fo-06 | Print Cartographer | Gallery-quality map prints with annotation, legend, compass rose |
| fo-07 | Timeline Weaver | Zoomable chronological narrative from temporal data |
| fo-08 | Projection Canvas | Real-time data to projection mapping, LED walls, Art-Net |
| fo-09 | Share Composer | Social and collaborative formatting with privacy gating |
| fo-10 | Voice Renderer | Text-to-speech from narrated prose—audio digests, walking tours |
Narrator
Transforms structured rill output into human-readable prose at three carefully tuned depth registers. Chroma delivers punchy, color-saturated summaries—a weather briefing you’d enjoy reading aloud. Lux provides measured, detailed narrative suitable for reports and field notes. Nocturne opens into contemplative long-form: the kind of writing that makes you pause and think. Every piece carries the Slipstream voice—warm, precise, grounded in place.
Dataviz Engine
The shared visualization library behind every chart, gauge, and heatmap in Slipstream. Dataviz Engine accepts normalized data arrays and renders them into responsive, accessible visual components—sparklines that breathe with real-time values, radar charts that overlay multiple rill outputs, heatmaps that reveal spatial patterns at a glance. It speaks the same visual language whether the data comes from weather, finance, health, or soil sensors.
Report Builder
Assembles multi-section documents from composed rill outputs, producing publication-ready PDFs, editable DOCX files, or responsive HTML pages. Report Builder handles page layout, section ordering, table of contents generation, header/footer templating, and figure placement. Feed it a narrative from fo-01, charts from fo-02, and formatting rules from stilus—it returns a document you’d be comfortable handing to a client or filing with a regulatory body.
Dashboard Composer
The bento-box interface where rills become tiles. Dashboard Composer lets users drag and drop rill outputs into responsive grid layouts, resize panels, set refresh intervals, and save named configurations. Each tile is a live window into a rill’s output—a weather gauge beside a soil moisture sparkline beside a market radar chart—all updating in concert. Saved dashboards become shareable, embeddable, and printable.
Field Guide Generator
Produces printable, pocket-sized guides tailored to a specific location, season, and activity. Ask for a foraging guide for the Olympic Peninsula in October and it pulls from botanic rills, terrain data, weather forecasts, and tidal schedules—then formats everything into a fold-friendly layout with illustrations, safety notes, and GPS waypoints. The kind of document you laminate and stuff in a daypack.
Print Cartographer
Renders gallery-quality map prints from geospatial rill data. Print Cartographer handles projection selection, scale bars, compass roses, legend composition, and annotation layers—producing maps that work as both analytical tools and wall art. Feed it terrain elevation, watershed boundaries, or species distribution data and it returns a map beautiful enough to frame, precise enough to navigate by.
Timeline Weaver
Transforms temporal data into zoomable, scrollable chronological narratives. Timeline Weaver handles multiple parallel tracks—overlay geological epochs with cultural events with personal milestones—and lets users zoom from millennia to minutes. Each event node can expand into detail cards with prose from fo-01, images, and links. The result is less a chart and more a story told through time.
Projection Canvas
Bridges the gap between data and physical space. Projection Canvas takes real-time rill outputs and maps them onto projection surfaces, LED walls, and Art-Net DMX universes. A weather rill becomes a living light installation; soil moisture data drives color gradients across a gallery wall; tidal data pulses through architectural lighting. It speaks Syphon, NDI, and Art-Net natively, making Slipstream data tangible in built environments.
Share Composer
Formats rill outputs for sharing across social platforms, collaborative workspaces, and messaging channels—with privacy gating baked in. Share Composer knows the constraints of each destination: Twitter’s character limits, Slack’s block kit, email’s HTML quirks, Open Graph for link previews. It strips sensitive fields, applies audience-appropriate detail levels, and generates shareable URLs with optional expiration and access controls.
Voice Renderer
Converts narrated prose from fo-01 into spoken audio—morning weather briefings you listen to while making coffee, botanical walking tours that narrate as you hike, bedtime summaries of the day’s data. Voice Renderer handles text-to-speech with configurable voice, pacing, and emphasis markers. It produces MP3/AAC files or real-time audio streams, complete with chapter markers and metadata for podcast players.
Domain VII: Machina
System Operations — How does the engine run?
Machina is the engine room. Every other domain produces intelligence about the world; Machina ensures that intelligence flows reliably, securely, and on time. It manages databases, guards authentication, routes navigation, schedules alerts, monitors system health, bridges external services, onboards new users, and orchestrates the temporal infrastructure that gives Slipstream its sense of time. Thirteen families, each responsible for one dimension of the operating layer that makes everything else possible.
datum (that which is given)
Datum is the plumbing beneath every rill. Before any intelligence can flow, data must be fetched, validated, cached, versioned, and backed up. This family manages the full lifecycle of data infrastructure—from database connectors and ETL pipelines to schema migrations and quality monitoring. It doesn’t care what the data means; it cares that the data arrives intact, on time, and in the shape downstream rills expect.
dt-01 → dt-05 → dt-08 → aer-01
Database connector fetches raw weather data, cache engine stores it locally, data quality validates the payload, then the aer rill transforms it into a forecast.
| ID | Name | One-liner |
|---|---|---|
| dt-01 | Database Connector | Supabase, PostgreSQL, SQLite connection pooling and management |
| dt-02 | ETL Pipeline | Extract, transform, load workflows with scheduling and retry |
| dt-03 | Schema Manager | Database schema versioning, diff, and migration generation |
| dt-04 | Data Sync | Cross-source synchronization with conflict resolution |
| dt-05 | Cache Engine | Multi-tier caching—memory, disk, CDN—with TTL and invalidation |
| dt-06 | Backup Manager | Automated backup scheduling, verification, and point-in-time restore |
| dt-07 | Query Optimizer | Query performance analysis, indexing recommendations, slow-query alerts |
| dt-08 | Data Quality | Validation rules, deduplication, integrity checks, freshness scoring |
| dt-09 | Migration Runner | Schema and data migration execution with rollback safety |
| dt-10 | Data Catalog | Metadata registry for all sources—lineage, ownership, refresh schedule |
Database Connector
The universal adapter between Slipstream and its data stores. Database Connector manages connection pools to Supabase (primary), PostgreSQL (analytics), and SQLite (local-first offline cache). It handles connection lifecycle, health checks, automatic reconnection, and credential rotation—ensuring every rill that needs data can get it without worrying about the plumbing underneath.
ETL Pipeline
Orchestrates the flow of data from external sources into Slipstream’s normalized format. ETL Pipeline defines extract steps (API calls, file reads, scrapes), transform steps (parsing, normalization, unit conversion), and load steps (database writes, cache updates). Each pipeline runs on a configurable schedule with retry logic, dead-letter queues for failed records, and detailed execution logs.
Schema Manager
Tracks the shape of every table, view, and index in Slipstream’s data stores. Schema Manager generates migration files from schema diffs, validates that proposed changes won’t break downstream rills, and maintains a version history so you can always answer the question: “What did the database look like on Tuesday?” It enforces naming conventions, foreign key integrity, and type safety across the entire data layer.
Data Sync
Keeps data consistent across Slipstream’s distributed stores—cloud database, local SQLite cache, edge CDN, and offline-first PWA storage. Data Sync implements conflict resolution strategies (last-write-wins, merge, manual), handles network partitions gracefully, and ensures that a user who goes offline in the field comes back online without losing observations. It’s the glue between cloud-first and local-first.
Cache Engine
Multi-tier caching that keeps Slipstream fast without serving stale data. Cache Engine manages three layers: in-memory (sub-millisecond, volatile), disk (persistent, larger), and CDN edge (global, read-heavy). Each cached item carries a TTL, staleness policy, and invalidation hooks—so when a weather API returns fresh data, every downstream rill that used the old forecast gets notified automatically.
Backup Manager
Automated backup scheduling with verification and point-in-time restore. Backup Manager runs daily incrementals and weekly full snapshots, verifies backup integrity via checksum comparison, and supports granular restore—recover a single table, a specific user’s data, or the entire database to any point in the last 90 days. It also manages backup rotation, off-site replication, and retention policies.
Query Optimizer
Watches every database query Slipstream executes and surfaces the ones that need attention. Query Optimizer profiles execution plans, identifies missing indexes, flags N+1 patterns, and recommends query rewrites—all automatically. When a rill’s data fetch slows from 50 ms to 500 ms because a table grew, this rill notices before users do and suggests the fix.
Data Quality
The immune system for Slipstream’s data layer. Data Quality runs validation rules against every incoming payload—type checking, range validation, null detection, duplicate identification, and schema conformance. It assigns a freshness score to every data source (how old is this reading?) and a confidence score (how reliable is this source historically?). When quality drops below threshold, downstream rills are warned before they produce misleading output.
Migration Runner
Executes the migration files that dt-03 generates. Migration Runner handles the careful, transactional application of schema changes—adding columns, creating indexes, backfilling data, renaming fields—with automatic rollback if anything fails mid-flight. It maintains a migration ledger so the system always knows exactly which migrations have been applied and which are pending.
Data Catalog
The metadata registry that answers “where does this data come from, who owns it, and when was it last refreshed?” Data Catalog tracks every data source in Slipstream—external APIs, user uploads, derived tables, cached computations—with lineage graphs showing how data flows from source to rill to output. It’s the map of the plumbing, making the invisible infrastructure visible and auditable.
securitas (freedom from care, safety)
Securitas guards the gates. Every API call, every user session, every data export passes through this family’s rills. It manages identity and authentication, encrypts data at rest and in transit, enforces role-based access, maintains an immutable audit trail, throttles abuse, and ensures privacy compliance. Security isn’t a feature bolted on at the end—it’s a family that wraps around everything else.
sc-01 → sc-03 → sc-04 → any-rill
Identity verifies the user, access control checks rill-level permissions, audit trail logs the access, then the requested rill executes.
| ID | Name | One-liner |
|---|---|---|
| sc-01 | Identity & Auth | User identity, OAuth providers, session management |
| sc-02 | Encryption Layer | Data encryption at rest and in transit, key management |
| sc-03 | Access Control | Per-rill permissions, RBAC, sharing controls, data sovereignty |
| sc-04 | Audit Trail | Immutable log of data access, modifications, sharing events |
| sc-05 | Rate Limiter | API usage throttling, abuse prevention, cost protection |
| sc-06 | Privacy Engine | Data minimization, anonymization, GDPR/CCPA compliance |
| sc-07 | Session Manager | JWT lifecycle, refresh tokens, concurrent session control |
| sc-08 | Two-Factor Auth | TOTP, WebAuthn, backup codes, recovery flows |
| sc-09 | Secret Vault | API key storage, rotation schedules, access logging |
| sc-10 | Threat Monitor | Brute force detection, suspicious activity, IP reputation scoring |
Identity & Auth
The front door. Identity & Auth manages the full authentication lifecycle—sign-up, sign-in, password reset, OAuth flows (Google, GitHub, Apple), magic links, and session establishment. It produces a verified identity token that every other rill can trust without re-authenticating. Built on Supabase Auth with custom claims for rill-level permissions.
Encryption Layer
Manages cryptographic operations across Slipstream—encrypting sensitive data at rest (user health records, API keys, personal observations), securing data in transit (TLS certificate management, certificate pinning), and handling key lifecycle (generation, rotation, revocation). Uses Web Crypto API for browser-side operations and libsodium for server-side, ensuring end-to-end encryption where privacy demands it.
Access Control
Determines who can do what with which rill. Access Control implements role-based access (RBAC) with granular per-rill permissions—read, write, share, export, delete—and supports data sovereignty controls that let users specify where their data lives and who can see it. Community features get separate permission layers: a shared observation is public, but the observer’s health data stays private.
Audit Trail
An append-only, tamper-evident log of every meaningful action in Slipstream. Audit Trail records who accessed what data, when, from where, and why—authentication events, permission changes, data exports, sharing actions, configuration modifications. The log is immutable: entries can be added but never edited or deleted. Essential for compliance, debugging, and answering the question “what happened?”
Rate Limiter
Protects Slipstream from accidental self-DDoS and external abuse. Rate Limiter enforces per-user, per-rill, and per-API quotas using a sliding window algorithm. It tracks API call budgets against external services (some of which charge per call), prevents runaway automation from burning through rate limits, and returns clear, actionable feedback when limits are hit—not just “429 Too Many Requests” but “you’ve used 80% of your OpenWeather quota; next refresh in 12 minutes.”
Privacy Engine
Enforces data minimization, anonymization, and regulatory compliance across every data flow. Privacy Engine strips personally identifiable information before community sharing, implements right-to-erasure (GDPR Article 17), manages consent records, and ensures data exports contain only what the user approved. It turns the abstract promise of “we respect your privacy” into concrete, auditable operations.
Session Manager
Manages the lifecycle of authenticated sessions after sc-01 verifies identity. Session Manager handles JWT issuance and refresh, tracks concurrent sessions across devices, enforces session timeouts, and supports instant revocation—log out everywhere with one click. It balances security (short token lifetimes) with usability (seamless refresh so users don’t get logged out mid-observation).
Two-Factor Auth
Adds a second verification layer beyond passwords. Two-Factor Auth supports TOTP (authenticator apps like Authy, Google Authenticator), WebAuthn (hardware keys like YubiKey, platform biometrics), and one-time backup codes for account recovery. It integrates with sc-01 as an optional-but-encouraged step, with gentle nudges rather than forced enrollment.
Secret Vault
Secure storage for the credentials Slipstream needs to operate—API keys, OAuth client secrets, database passwords, webhook signing keys. Secret Vault encrypts everything at rest via sc-02, enforces rotation schedules (auto-rotate every 90 days, alert on stale keys), logs every access via sc-04, and provides a clean interface so no rill ever has a hardcoded secret in its configuration.
Threat Monitor
Watches for patterns that suggest something is wrong—brute force login attempts, credential stuffing, session hijacking, unusual geographic access patterns, API scraping. Threat Monitor correlates signals from sc-01, sc-04, and sc-05 to build a real-time threat picture, automatically escalating from warnings to temporary IP blocks to full account lockouts based on severity. It’s the night watch.
regia (royal road, guidance)
Regia is the user’s map through Slipstream. It manages everything the user touches to navigate: the home dashboard, workspace switching, search, favorites, notification center, command palette, and keyboard shortcuts. Without regia, 400+ rills would be an overwhelming wall of capability. With it, each user sees a personalized, context-aware surface that puts the right rill at their fingertips at the right moment.
rg-02 → rg-03 → fo-04 → rg-01
User switches to “Garden” workspace, favorites load, dashboard composer arranges pinned rills, home dashboard renders the personalized view.
| ID | Name | One-liner |
|---|---|---|
| rg-01 | Home Dashboard | Primary landing page—personalized, time-aware, context-rich |
| rg-02 | Workspaces | Named contexts: “Garden,” “Beekeeping,” “Health”—switchable |
| rg-03 | Favorites | Pinned rills and flows for quick access across workspaces |
| rg-04 | Search | Full-text search across all rill outputs, journals, community data |
| rg-05 | Explore | Browse all rills by family, domain, or interest with live previews |
| rg-06 | Notifications Center | Unified inbox for alerts, digests, and community messages |
| rg-07 | History | Timeline of all interactions, observations, queries, and changes |
| rg-08 | Multi-Location | Location manager with auto-recontextualization when you move |
| rg-09 | Command Palette | Keyboard-driven command interface for power users |
| rg-10 | Keyboard Shortcuts | Customizable hotkeys for all major navigation and rill actions |
Home Dashboard
The first thing you see when you open Slipstream. Home Dashboard assembles a personalized, time-aware landing page—morning shows weather, circadian state, and today’s tasks; evening shifts to a day-in-review summary with health metrics and tomorrow’s forecast. It draws from the user’s active workspace, favorites, and recent activity to surface the most relevant rills without overwhelming with options.
Workspaces
Named contexts that filter the entire Slipstream experience. Switch to “Garden” and soil, weather, phenology, and task rills come to the foreground; switch to “Health” and circadian, nutrition, sleep, and movement rills take over. Workspaces are user-defined bundles of rill subscriptions, dashboard layouts, and notification preferences—one person’s Slipstream can feel completely different from another’s.
Favorites
Quick-access pins for the rills and flows you use most. Favorites maintains an ordered list of bookmarked items that appear in the sidebar, home dashboard, and command palette results. They can be workspace-scoped (only visible in “Garden”) or global (visible everywhere). Drag to reorder, group into folders, and share a favorites bundle with another user to help them get started.
Search
Full-text search across everything in Slipstream—rill outputs, journal entries, community observations, dashboard configurations, and documentation. Search indexes content as it’s produced, supports fuzzy matching and filters (by domain, family, date range, data type), and ranks results by relevance and recency. Type “soil moisture last week” and get the specific readings, the trend chart, and the journal entry where you noted the irrigation change.
Explore
The browsable catalog of every rill in Slipstream, organized by domain, family, and interest area. Explore shows live previews of rill output using your current location and data, so you can see what a rill does before activating it. Filter by tag, sort by popularity or relevance, and discover compositions—rills that work better together than alone.
Notifications Center
The unified inbox for everything that wants your attention—weather alerts, task reminders, community messages, system health warnings, and digest summaries. Notifications Center categorizes, prioritizes, and groups related items so you see “3 garden alerts” instead of three separate interruptions. Integrates with nu-02 for routing and nu-05 for intelligent timing.
History
A scrollable timeline of every interaction with Slipstream—rill views, searches, journal entries, configuration changes, shared items, and system events. History lets you revisit what you were looking at last Tuesday, trace how your usage patterns evolved over months, and recover context after a break. Filterable by type, searchable by content, and exportable for personal data portability.
Multi-Location
For people who live in more than one place—home, office, cabin, garden plot, parents’ house. Multi-Location manages named locations with coordinates, timezone, and associated rills, and auto-recontextualizes when you arrive: weather switches to local forecast, soil data to the right garden, community feed to the nearest circle. Switch manually or let geofencing handle it.
Command Palette
Press Cmd+K and type anything. Command Palette is the keyboard-driven interface for power users who want to navigate, search, execute actions, and switch contexts without touching the mouse. It indexes every rill, every workspace, every action, and every setting into a fuzzy-searchable command list. Built on cmdk, it’s the fastest way to do anything in Slipstream.
Keyboard Shortcuts
Customizable hotkeys for every major action in Slipstream. Keyboard Shortcuts manages a global shortcut registry, handles conflicts between rill-specific and global bindings, supports multi-key chords, and provides a visual reference sheet (press ? to see all shortcuts). Defaults follow conventions from VS Code and Figma; everything is remappable.
praxis (practice, action, doing)
Praxis is how Slipstream learns your preferences and adapts to your life. It manages identity profiles, rill subscriptions, privacy controls, display themes, location contexts, API key management, integration settings, feature flags, experimentation, and configuration portability. Where regia handles navigation, praxis handles personalization—making the system feel like yours.
px-01 → px-05 → rg-08 → aer-01
Identity profile provides chronotype, location manager sets active coordinates, multi-location applies context, weather rill delivers a forecast tuned to this person, this place, this time.
| ID | Name | One-liner |
|---|---|---|
| px-01 | Identity Profile | Progressive self-description: location, interests, chronotype, birth data |
| px-02 | Subscriptions | Which rills are active, which alerts enabled, threshold settings |
| px-03 | Data Sovereignty | Per-rill privacy controls, sharing permissions, export/delete |
| px-04 | Theme & Display | Nocturne/Lux/Chroma mode, density, accessibility settings |
| px-05 | Location Manager | Multi-place living with switchable context and default coordinates |
| px-06 | API Key Manager | User-provided API keys, usage tracking, cost limits |
| px-07 | Integration Settings | Connected services: calendar, weather station, hive scale, etc. |
| px-08 | Feature Flags | Progressive feature rollout, beta access, canary deployments |
| px-09 | A/B Testing | Experiment framework with statistical significance tracking |
| px-10 | Config Export | Settings export/import, profile migration, backup and restore |
Identity Profile
A progressive self-description that grows as the user shares more about themselves. Start with just a location and interest; over time add chronotype (early bird, night owl), birth data (for celestial rills), dietary preferences (for nutrition rills), skin type (for botanical wellness). Nothing is required; everything is optional. The more context provided, the more personalized the intelligence becomes.
Subscriptions
Controls which rills are active for this user and how they behave. Subscriptions manages the on/off state of every rill, alert threshold settings (notify me when soil moisture drops below 30%), refresh intervals (check weather every 15 minutes), and notification preferences per rill. It’s the control panel for tuning Slipstream to your signal-to-noise ratio.
Data Sovereignty
Gives users granular control over their data—per-rill privacy settings, sharing permissions, data export, and the right to delete. Data Sovereignty answers: “Can this rill share my observations with the community? Can I export all my health data? Can I delete everything and start over?” The answer is always yes, and the controls are always accessible, never buried.
Theme & Display
Controls the visual personality of Slipstream. Theme & Display manages three core modes—Nocturne (dark, low-light), Lux (bright, high-contrast), and Chroma (vivid, color-saturated)—plus density settings (comfortable, compact), font size, motion preferences (reduce motion for accessibility), and color vision accommodations. The theme follows M3 Design System tokens via CSS custom properties.
Location Manager
Stores and manages the user’s named locations—home, office, garden, cabin—with coordinates, timezone, elevation, and associated sensors. Location Manager feeds rg-08 for multi-location switching and provides the geographic context that every location-aware rill depends on. Add a new location by dropping a pin, entering an address, or importing GPS coordinates.
API Key Manager
Users who want premium API access or bring their own keys manage them here. API Key Manager stores user-provided keys in sc-09 (Secret Vault), tracks per-key usage against cost limits, and provides clear visibility into spending: “Your OpenAI key has used $4.20 of your $10 monthly budget.” Keys are never exposed in the UI after initial entry—only masked previews and usage stats.
Integration Settings
Configuration hub for every external service Slipstream connects to—Google Calendar, personal weather stations, hive scales, iNaturalist, Home Assistant, and more. Integration Settings manages OAuth connections, polling intervals, data mapping, and health status for each integration. When a weather station goes offline, you see it here first.
Feature Flags
Controls progressive feature rollout across the Slipstream user base. Feature Flags manages boolean and multivariate flags that gate access to new rills, experimental UI, beta integrations, and infrastructure changes. Supports percentage rollouts (10% of users see the new dashboard), user targeting (beta testers get early access), and instant kill switches for anything that goes wrong.
A/B Testing
Experiment framework that lets Slipstream test hypotheses with real users. A/B Testing builds on px-08 feature flags to randomly assign users to experiment variants, tracks conversion metrics, and calculates statistical significance so decisions are data-driven. “Does the bento dashboard or the list view lead to more rill activations?”—run the experiment, measure, ship the winner.
Config Export
Export your entire Slipstream configuration—profile, subscriptions, themes, integration settings, favorites, workspace layouts—as a portable JSON file. Import it on a new device, share it with a friend to help them set up, or keep it as a backup. Config Export also handles migration between Slipstream versions, mapping old config schemas to new ones without losing settings.
kairos (the opportune moment)
Kairos is the event nervous system. It watches every rill’s output for conditions that matter—anomalies, threshold crossings, converging signals—and orchestrates the response: routing alerts, triggering cascades, scheduling recurring checks, managing incident escalation. Where content rills detect what’s happening, kairos decides who needs to know and how urgently.
aer-01 → kr-01 → kr-02 → kr-08 → nu-03
Weather rill detects anomaly, anomaly detector flags it, threshold engine confirms it crosses user limits, alert composer formats the message, channel bridge delivers via push notification.
| ID | Name | One-liner |
|---|---|---|
| kr-01 | Anomaly Detector | Statistical outlier detection across all rill outputs |
| kr-02 | Threshold Engine | User-defined alert thresholds with hysteresis and debounce |
| kr-03 | Webhook Manager | Inbound/outbound webhook registration, dispatch, and retry |
| kr-04 | Event Cascade | Multi-step event chains with conditional branching |
| kr-05 | Cron Scheduler | Time-based task scheduling with timezone awareness |
| kr-06 | Incident Router | Alert escalation, acknowledgment tracking, on-call routing |
| kr-07 | Event Log | Structured event stream with retention and replay capability |
| kr-08 | Alert Composer | Multi-channel alert formatting for push, email, SMS, in-app |
| kr-09 | Cooldown Manager | Alert fatigue prevention—deduplication, batching, suppression |
| kr-10 | Event Replay | Historical event playback for debugging and analysis |
Anomaly Detector
Watches every rill’s output for values that don’t fit the pattern. Anomaly Detector applies statistical methods—z-scores for normally distributed data, IQR for skewed distributions, seasonal decomposition for time series—to flag readings that are genuinely unusual, not just noisy. A temperature of 45°F in July in Portland is an anomaly; 45°F in January is Tuesday.
Threshold Engine
Evaluates user-defined alert rules against live rill data. Threshold Engine supports simple comparisons (“soil moisture < 30%”), compound conditions (“temperature > 90°F AND humidity > 80%”), and hysteresis to prevent alert flapping—once triggered, a threshold doesn’t clear until the value recovers past a reset point, preventing the maddening on/off/on/off of a value hovering near the boundary.
Webhook Manager
Manages inbound and outbound webhook connections. Inbound: register a URL that external services can POST to, triggering rill actions when events arrive from GitHub, Stripe, weather stations, or IoT devices. Outbound: send rill events to external URLs with configurable payloads, retry logic, and signing keys for verification. Webhook Manager is the generic event bridge between Slipstream and the rest of the internet.
Event Cascade
Orchestrates multi-step event chains where one event triggers the next. Event Cascade defines workflows: “When frost alert fires, check greenhouse heater status, if off then send push notification AND create task to cover outdoor plants.” Supports conditional branching, parallel execution, delay steps, and circuit breakers to prevent runaway cascades.
Cron Scheduler
Time-based scheduling for recurring tasks, data refreshes, and periodic checks. Cron Scheduler supports standard cron expressions with timezone awareness—so “every morning at 6 AM” means 6 AM in your timezone, not UTC. It manages the scheduling of ETL pipelines, digest generation, backup runs, and any rill that needs to execute on a clock rather than on-demand.
Incident Router
Escalation engine for alerts that need human attention. Incident Router manages severity levels (info → warning → critical → emergency), acknowledgment tracking (“has anyone seen this?”), and escalation policies (“if no acknowledgment in 15 minutes, escalate to push notification; if still no response in 30 minutes, call”). For multi-user setups, it supports on-call rotation.
Event Log
Structured, append-only stream of every event that flows through kairos. Event Log captures alerts, threshold crossings, cascade executions, webhook deliveries, and cron runs with consistent schema—making it possible to answer “what happened in my system between 2 AM and 4 AM last night?” Supports retention policies, log rotation, and export for external analysis.
Alert Composer
Formats alert content for each delivery channel. Alert Composer takes a raw alert event and produces channel-appropriate versions: a concise push notification (under 100 chars), a detailed email with charts and context, an SMS with just the critical facts, and an in-app card with action buttons. Same alert, four formats—each optimized for its medium.
Cooldown Manager
Prevents alert fatigue by managing suppression, deduplication, and intelligent batching. Cooldown Manager enforces minimum intervals between repeated alerts (“don’t tell me about soil moisture more than once per hour”), deduplicates identical events, and batches low-priority alerts into digests rather than individual notifications. The goal: every alert that reaches you is worth reading.
Event Replay
Replays historical events through the kairos pipeline for debugging, testing, and analysis. Event Replay lets you re-run last Tuesday’s events through updated threshold rules to see what would have triggered, test new cascade workflows against real event data, and investigate incidents by watching the event stream unfold in slow motion. Think of it as a DVR for system events.
census (assessment, registration)
Census watches the watchers. While every other family monitors the external world, census monitors Slipstream itself—rill health, API uptime, data freshness, usage patterns, error rates, performance metrics, and cost tracking. It’s the system’s self-awareness layer, ensuring that when a data source goes down or a rill starts returning stale results, you know about it before the outputs become unreliable.
ce-01 → ce-03 → kr-01 → kr-06 → nu-03
Rill health detects an API failure, data freshness flags stale results, anomaly detector confirms it’s outside normal, incident router escalates, and the user gets a push notification.
| ID | Name | One-liner |
|---|---|---|
| ce-01 | Rill Health | Per-rill status: API up/down, last fetch, error rate, latency |
| ce-02 | Usage Insights | Personal analytics: which rills you use most, when, patterns |
| ce-03 | Data Freshness | How old is the data behind each rill? Transparency dashboard |
| ce-04 | Cost Tracker | API usage, estimated cost, budget alerts for paid services |
| ce-05 | Error Tracker | Centralized error capture, deduplication, and trend analysis |
| ce-06 | Performance Monitor | Web Vitals, rill execution time, rendering performance |
| ce-07 | Uptime Monitor | Continuous availability checks for all external data sources |
| ce-08 | Dependency Graph | Visual map of rill-to-rill and rill-to-API dependencies |
| ce-09 | Capacity Planner | Storage usage, growth projections, quota warnings |
| ce-10 | System Dashboard | Unified ops view: health, costs, errors, performance at a glance |
Rill Health
Monitors the operational status of every rill in Slipstream. Rill Health tracks per-rill metrics: is the upstream API responding? When was the last successful data fetch? What’s the error rate over the last hour? What’s the p95 latency? It produces a simple stoplight status (green/yellow/red) for each rill plus detailed diagnostics for troubleshooting when things go wrong.
Usage Insights
Personal analytics about how you use Slipstream. Usage Insights tracks which rills you visit most, when you’re most active (morning weather check, evening health review), which compositions you run frequently, and how your usage evolves over time. It’s not surveillance—it’s self-knowledge, helping you understand your own patterns and discover rills you might be overlooking.
Data Freshness
Answers the question every user should be able to ask: “How old is this data?” Data Freshness tracks the last-updated timestamp for every data source behind every rill, calculates staleness relative to expected refresh intervals, and surfaces transparency indicators in the UI. A weather forecast from 6 hours ago gets a yellow badge; one from yesterday gets red. Trust requires transparency.
Cost Tracker
Tracks API usage costs across all external services Slipstream consumes. Cost Tracker tallies calls to paid APIs (weather, AI models, satellite imagery), estimates monthly spending, and triggers budget alerts before you hit limits. It distinguishes between Slipstream-provided API access (included in subscription) and user-provided keys (tracked against their personal budgets in px-06).
Error Tracker
Centralized error capture for the entire Slipstream stack. Error Tracker collects exceptions, API failures, timeout errors, and validation failures from all rills and infrastructure, deduplicates them by root cause, tracks frequency trends, and surfaces the errors that matter most. Integration with Sentry for stack traces and source maps; custom grouping logic for rill-specific error patterns.
Performance Monitor
Tracks the speed and responsiveness of every surface in Slipstream. Performance Monitor collects Core Web Vitals (LCP, FID, CLS), rill execution times, rendering durations, and network waterfalls. It identifies performance regressions before users complain: “The weather rill went from 200 ms to 1.2 s after yesterday’s deploy”—and suggests where the bottleneck lives.
Uptime Monitor
Continuous availability checks for every external data source Slipstream depends on. Uptime Monitor pings APIs at regular intervals, tracks uptime percentages, measures response times, and maintains a historical availability record. When OpenWeather goes down, Slipstream knows within 5 minutes and can show users a “data source temporarily unavailable” notice instead of cryptic errors.
Dependency Graph
A visual, interactive map of how everything in Slipstream connects. Dependency Graph shows rill-to-rill compositions, rill-to-API dependencies, and data flow paths—answering questions like “If the USDA API goes down, which rills are affected?” or “What upstream rills feed into my morning digest?” Click any node to see its health status, freshness, and downstream impact.
Capacity Planner
Tracks storage usage, database size, cache utilization, and growth trends to prevent surprise quota exhaustion. Capacity Planner projects when you’ll hit limits based on current growth rates, recommends cleanup actions (archive old data, prune stale caches), and alerts well before hard limits are reached. It also tracks Supabase plan limits and Vercel bandwidth.
System Dashboard
The unified ops view that pulls together everything census knows into a single screen. System Dashboard shows rill health status (ce-01), data freshness (ce-03), cost summary (ce-04), error trends (ce-05), performance metrics (ce-06), uptime status (ce-07), and capacity projections (ce-09)—all at a glance. It’s the cockpit for anyone who wants to know “is Slipstream healthy right now?”
nuntius (messenger, envoy)
Nuntius is the postal service. Where kairos decides what to communicate and how urgently, nuntius handles how it gets delivered—composing morning digests, routing alerts to the right channel (email, push, SMS, Slack), learning when you want to be interrupted and when you don’t, and producing periodic summaries that turn raw data into narrative. The messenger should be as thoughtful as the message.
aer-01 + fl-01 + sa-01 → nu-01 → nu-05 → nu-03
Weather, flora, and health rills feed the digest composer, quiet intelligence decides this is a good moment to deliver, channel bridge sends it as a morning email.
| ID | Name | One-liner |
|---|---|---|
| nu-01 | Digest Composer | Multi-rill morning briefing assembled into narrative |
| nu-02 | Alert Router | Urgency scoring, batching, quiet hours, notification management |
| nu-03 | Channel Bridge | Delivery to email, push notification, SMS, Slack, Discord |
| nu-04 | Voice Briefing | Audio version of the daily digest for hands-free listening |
| nu-05 | Quiet Intelligence | Learns your patterns, holds low-priority items until you’re ready |
| nu-06 | Seasonal Summary | Weekly/monthly/seasonal retrospective narrative |
| nu-07 | Delivery Analytics | Open rates, click rates, delivery failures, channel effectiveness |
| nu-08 | Schedule Manager | Delivery timing: digest schedule, timezone-aware send windows |
Digest Composer
Assembles a multi-rill morning briefing into a coherent narrative—not a list of data points, but a story about your day. Digest Composer pulls from weather, health, garden, financial, and task rills (based on your subscriptions), feeds them through fo-01 (Narrator), and produces a briefing that reads like a letter from a well-informed friend. “Good morning. It’s 52° and clearing—perfect for the garden work you planned. Your soil moisture is holding at 45%, so skip watering today.”
Alert Router
The traffic controller between kairos (which generates alerts) and nu-03 (which delivers them). Alert Router scores urgency, respects quiet hours, batches low-priority items, and decides which channel is appropriate—critical frost warnings go to push notification immediately; low-priority usage insights wait for the weekly digest. It’s the intelligence layer between “something happened” and “someone got interrupted.”
Channel Bridge
The last mile—actually sending the notification through the right channel. Channel Bridge integrates with email providers (Resend, SendGrid), push notification services (web-push, Firebase Cloud Messaging), SMS gateways, Slack webhooks, and Discord bots. Each channel gets content formatted by kr-08 (Alert Composer) and timing from nu-02. It handles delivery confirmation, bounce processing, and retry logic.
Voice Briefing
The audio edition of your morning digest. Voice Briefing takes the narrative from nu-01, runs it through fo-10 (Voice Renderer), and produces a 2–5 minute spoken briefing you can listen to while making coffee, driving, or walking the dog. “Hey Slipstream, what’s happening?”—and it tells you, in a warm, unhurried voice, everything that matters right now.
Quiet Intelligence
Learns when you want to be interrupted and when you don’t. Quiet Intelligence observes your interaction patterns—when you read notifications, when you dismiss them, when you’re active in the app—and builds a model of your attention rhythms. Low-priority items are held until the right moment; urgent items break through any quiet period. The goal: every notification arrives when you’re ready to receive it.
Seasonal Summary
Periodic retrospective narratives that step back from the daily stream and look at the bigger picture. Seasonal Summary generates weekly recaps, monthly reviews, and seasonal retrospectives—how your garden evolved this spring, how your health metrics trended this quarter, what the weather pattern looked like over the last season. It’s the chapter-level view of your data story.
Delivery Analytics
Tracks how effectively notifications are reaching and engaging users. Delivery Analytics measures open rates, click-through rates, delivery failures, bounce rates, and channel effectiveness—so Slipstream can learn which channels work best for which types of messages. If email digests have 80% open rates but push notifications get dismissed 90% of the time, the system adapts.
Schedule Manager
Controls the timing of all scheduled deliveries—when the morning digest sends, what timezone it uses, which days get weekly summaries, and how seasonal summaries align with astronomical events. Schedule Manager respects time zones, handles daylight saving transitions, and lets users set send windows: “No notifications before 7 AM or after 10 PM, except critical alerts.”
retia (nets, networks)
Retia treats the network as infrastructure with its own topology, physics, and failure modes. This family discovers devices and services on local networks, maps WiFi signal coverage, monitors bandwidth, manages port exposure, and—at its most advanced—uses WiFi Channel State Information for presence detection and environmental sensing. Three core principles: passive first, radio waves are sensors, and a tiered ethical framework governs access.
rt-01 → rt-02 → rt-05 → ce-08
Topology mapper discovers network shape, service beacon finds active services, bandwidth sentinel measures throughput, dependency graph integrates network status into the system health view.
| ID | Name | One-liner |
|---|---|---|
| rt-01 | Topology Mapper | Network topology discovery, visualization, and change detection |
| rt-02 | Service Beacon | Local service discovery, registration, and health monitoring |
| rt-03 | Signal Cartographer | WiFi signal strength mapping, coverage visualization, dead zones |
| rt-04 | Local Vault | Local network storage discovery, configuration, and access |
| rt-05 | Bandwidth Sentinel | Throughput monitoring, traffic classification, QoS awareness |
| rt-06 | Port Warden | Port management, firewall awareness, service exposure control |
| rt-07 | Aether Sense | WiFi CSI presence and motion detection (Tier 2: Authenticated) |
| rt-08 | Aether Scan | Advanced WiFi CSI spatial awareness and environmental sensing |
| rt-09 | Shield Audit | Network security posture assessment and vulnerability scanning |
| rt-10 | Mesh Architect | Network planning, VLAN design, segmentation, optimization |
Topology Mapper
Discovers and visualizes the shape of your local network—routers, switches, access points, connected devices—using passive ARP observation and active probing. Topology Mapper detects changes (new device appeared, printer went offline) and maintains a living map that updates in real time. It’s the foundation for every other retia rill: you can’t manage what you can’t see.
Service Beacon
Discovers and monitors services running on the local network—Home Assistant instances, NAS shares, media servers, IoT hubs, printers. Service Beacon uses mDNS/Bonjour and SSDP/UPnP for zero-config discovery, then maintains health checks on found services. When your Home Assistant goes offline or your NAS runs low on storage, Service Beacon is the first to know.
Signal Cartographer
Maps WiFi signal strength across physical space, creating coverage heatmaps that reveal dead zones, interference patterns, and optimal access point placement. Walk through your home or office with a device running Signal Cartographer and it builds a room-by-room picture of where WiFi is strong, weak, or competing with neighboring networks. The map updates as conditions change.
Local Vault
Discovers and manages network-attached storage on the local network—NAS devices, shared drives, Time Machine targets, and SMB/NFS shares. Local Vault monitors storage capacity, health status (SMART data where available), and access permissions. It provides a unified view of all local storage alongside Slipstream’s cloud storage for a complete picture of where data lives.
Bandwidth Sentinel
Monitors network throughput, classifies traffic by type (streaming, browsing, IoT, backup), and provides quality-of-service awareness. Bandwidth Sentinel tracks upload/download speeds over time, identifies bandwidth hogs, and alerts when network performance degrades. It helps answer: “Why is everything slow right now?” and “What’s consuming all the bandwidth?”
Port Warden
Monitors which ports are open on local devices, what services they expose, and whether the exposure is intentional. Port Warden maps listening ports, compares against expected service configurations, and alerts when unexpected ports appear—a sign of misconfiguration or compromise. It’s firewall awareness without firewall management: see what’s exposed, decide if that’s what you want.
Aether Sense
Uses WiFi Channel State Information (CSI) for passive presence and motion detection—no cameras, no wearables, just the radio waves already in your home. ESP32 sensors measure how WiFi signals scatter off moving bodies, detecting room occupancy, movement direction, and activity level. Tier 2 (Authenticated) access: requires explicit user setup and consent. Hardware cost: 1–3 ESP32 modules per zone ($5–$30).
Aether Scan
Advanced WiFi CSI analysis that goes beyond presence detection into spatial awareness and environmental sensing. Aether Scan can detect breathing patterns (for sleep monitoring without wearables), estimate room occupancy count, sense door/window open/close states through signal path changes, and detect environmental changes like water leaks via humidity-affected signal propagation. Tier 2 (Authenticated): requires setup, calibration, and explicit consent.
Shield Audit
Assesses the security posture of your local network. Shield Audit checks for common vulnerabilities—default passwords on routers, unencrypted services, outdated firmware, open admin panels, weak WiFi encryption—and produces a security scorecard with prioritized recommendations. Tier 3 (Verified): requires explicit user authorization because it performs active assessment of network assets.
Mesh Architect
Network planning and optimization intelligence. Mesh Architect takes the current topology (rt-01), signal map (rt-03), and bandwidth data (rt-05) and recommends improvements—optimal access point placement, VLAN segmentation for IoT devices, channel assignment to minimize interference, and network architecture patterns for different use cases (smart home, studio, workshop). Design the network before you buy the hardware.
tessera (token, mosaic piece)
Tessera—a piece of mosaic—fits Slipstream into the broader digital ecosystem without trying to replace existing services. Each rill wraps one external service’s API: calendar sync, weather station ingest, hive scale data, photo pipelines, document export, MCP bridge, webhooks, and home automation. The MCP Bridge (ts-06) is architecturally critical—it exposes Slipstream rills as tools for AI agents.
ts-02 → aer-01 → fo-01 → ts-01
Personal weather station data feeds the weather rill, narrator writes a prose forecast, calendar sync pushes a “frost tonight” event to Google Calendar.
| ID | Name | One-liner |
|---|---|---|
| ts-01 | Calendar Sync | Push Slipstream events to Google/Apple Calendar |
| ts-02 | Weather Station Ingest | Personal weather hardware: Davis, Ecowitt, Ambient Weather |
| ts-03 | Hive Scale Ingest | Smart hive scale integration: BroodMinder, Arnia, HiveWatch |
| ts-04 | Photo Pipeline | Photos auto-tagged with rill context: weather, species, location |
| ts-05 | Export Engine | PDF reports, slide decks, social posts, blog drafts from rill output |
| ts-06 | MCP Bridge | Expose Slipstream rills as MCP tools for AI agents |
| ts-07 | Webhook Gateway | Generic inbound/outbound webhooks for any service |
| ts-08 | Home Automation | Home Assistant, HomeKit, IFTTT—trigger home actions from rills |
| ts-09 | iNaturalist Bridge | Species observation sync with iNaturalist community science |
| ts-10 | eBird Bridge | Bird observation sync with eBird/Cornell Lab databases |
Calendar Sync
Pushes Slipstream events to your existing calendar—Google Calendar, Apple Calendar (CalDAV), or any ICS-compatible service. When Slipstream detects a frost warning, a peak bloom window, or an optimal foraging day, it creates a calendar event with rill-powered detail in the notes: temperature forecast, species to look for, gear to bring. Your calendar becomes rill-aware without leaving your existing workflow.
Weather Station Ingest
Pulls hyperlocal weather data from personal weather hardware—Davis Vantage Pro, Ecowitt GW1000, Ambient Weather WS-2000—giving Slipstream readings from your backyard, not the nearest airport 10 miles away. Weather Station Ingest normalizes data across manufacturer APIs into a common schema and feeds it to weather rills as a primary or supplementary data source, dramatically improving forecast accuracy for your specific microclimate.
Hive Scale Ingest
Connects smart hive scale hardware to Slipstream’s apiary intelligence. Hive Scale Ingest pulls real-time weight, temperature, and humidity data from BroodMinder, Arnia, or HiveWatch devices, normalizing readings across manufacturers. Weight changes tell the story: rapid gain means nectar flow; sudden loss means swarm. Combined with weather and floral rills, it becomes a complete hive health picture.
Photo Pipeline
Auto-tags photos with rill context at the moment of capture. Take a photo in your garden and Photo Pipeline enriches it with weather conditions, soil moisture, what’s in bloom, nearby species predictions, and hive status—all pulled from active rills at that location and time. Months later, the photo isn’t just a picture; it’s a data-rich field observation with environmental context embedded in EXIF and sidecar metadata.
Export Engine
Generates polished, shareable outputs from rill data—PDF reports for clients, slide decks for presentations, social media posts for sharing, and blog draft markdown for publishing. Export Engine bridges the gap between Slipstream’s internal data and the external formats people actually use to communicate. It works closely with fo-03 (Report Builder) and fo-09 (Share Composer) but focuses on cross-platform file generation.
MCP Bridge
The architecturally critical rill that exposes Slipstream as a Model Context Protocol server. MCP Bridge takes rill manifests—inputs, outputs, descriptions—and publishes them as MCP tool definitions that AI agents (Claude, GPT, custom agents) can discover and call. Ask Claude “what’s the weather at my garden?” and it calls Slipstream via MCP, getting real hyperlocal data instead of hallucinating. This is how Slipstream becomes infrastructure for AI, not just a UI.
Webhook Gateway
Generic webhook infrastructure for connecting Slipstream to any service that speaks HTTP. Webhook Gateway provides unique inbound URLs that external services can POST to (GitHub deploys, Stripe payments, IFTTT triggers), and outbound hooks that fire when rill conditions are met. It works with kr-03 (Webhook Manager) for event processing but provides the transport layer and URL management.
Home Automation
Connects Slipstream to home automation platforms—Home Assistant (REST API), Apple HomeKit (via HomeBridge), and IFTTT—so rill intelligence can trigger physical actions. Frost warning? Turn on the greenhouse heater. Sunset approaching and you’re home? Dim the lights to Nocturne mode. Air quality dropping? Start the air purifier. Data-driven home automation with Slipstream as the intelligence layer.
iNaturalist Bridge
Syncs species observations between Slipstream and iNaturalist—the world’s largest community science biodiversity platform. When you log a plant or animal observation in Slipstream, iNaturalist Bridge can push it to iNaturalist for community identification and scientific contribution. In return, it pulls nearby observations from iNaturalist to enrich Slipstream’s local biodiversity picture.
eBird Bridge
Connects Slipstream to Cornell Lab’s eBird—the world’s largest bird observation database. eBird Bridge pulls recent sightings, hotspot data, and species frequency charts for your area, and can push your bird observations to eBird checklists. It enriches Slipstream’s fauna rills with real-time community birding data: what’s been seen nearby, what’s rare, what’s migrating through.
initium (beginning, entrance)
Initium is the beginning—how new users find their way in and how existing users discover capabilities they didn’t know existed. The critical rill is in-01 (First Flow): enter your location and one interest, and within 60 seconds live rills are running with your data. Not a tutorial—a demonstration. The system shows itself working in real time, and the user decides how deep to go from there.
in-01 → px-01 → px-02 → rg-01
First Flow captures location and interest, identity profile stores them, subscriptions activate relevant rills, and the home dashboard renders a personalized view in under a minute.
| ID | Name | One-liner |
|---|---|---|
| in-01 | First Flow | 60-second onboarding: location + interest → live rills running |
| in-02 | Rill Recommendations | Personalized discovery based on interest graph and usage patterns |
| in-03 | Seasonal Spotlight | “It’s October—here’s what’s special right now” |
| in-04 | Use Case Stories | Short narratives showing how real people use Slipstream |
| in-05 | Guided Tours | Interactive walkthroughs of specific flows using your actual data |
| in-06 | What’s New | Contextualized changelog—feature releases relevant to your profile |
| in-07 | Gift a Flow | Pre-configured flow URLs that onboard recipients with zero friction |
| in-08 | Interest Calibrator | Progressive preference refinement through implicit and explicit signals |
| in-09 | Migration Assistant | Import data and settings from other platforms and services |
| in-10 | Achievement System | Milestones and discovery badges that encourage exploration |
First Flow
The most important 60 seconds in Slipstream. First Flow asks two things: where are you, and what interests you? Enter “Portland, OR” and “gardening” and within a minute, live rills are running—local weather, soil conditions, what’s in bloom, frost risk, phenology calendar—all populated with real data for your location. Not a promise of what the system could do. A demonstration of what it’s already doing, right now, for you.
Rill Recommendations
Personalized rill discovery that goes beyond “you might also like.” Rill Recommendations analyzes your interest graph, usage patterns from ce-02, and rill dependency chains to surface capabilities you’d benefit from but haven’t activated. “You use soil moisture and weather rills heavily—the phenology calendar would show you how they connect to bloom timing.” Context-aware suggestions, not generic upsells.
Seasonal Spotlight
Time-aware rill discovery that highlights what’s special right now. In October, Seasonal Spotlight surfaces mushroom foraging rills, fall color tracking, salmon run predictions, and harvest calendars. In January, it’s aurora probability, seed catalog season, and indoor gardening. Each spotlight is location-aware—October in Portland looks different from October in Austin.
Use Case Stories
Short narratives showing how real people use Slipstream—a beekeeper in Oregon tracking nectar flow, a gardener in Vermont planning by phenology, a health-conscious runner correlating sleep with air quality. Use Case Stories provide social proof and imagination fuel: “I didn’t know I could do that.” Each story includes the specific rill composition that makes it work, so users can replicate it in one click.
Guided Tours
Interactive walkthroughs that guide users through specific workflows using their actual data—not screenshots of demo data. Guided Tours highlight UI elements, explain what each rill is showing, and let users interact at each step. “This is your soil moisture reading. See how it correlates with yesterday’s rain? Click here to set an alert for when it drops below 30%.” Learning by doing, with guardrails.
What’s New
A contextualized changelog that surfaces feature releases relevant to the user’s profile and interests. Not a raw list of commits—a curated announcement that says “We added a mushroom foraging rill, and since you’re in the Pacific Northwest with a gardening workspace, you might want to activate it.” Unobtrusive, dismissable, and genuinely useful.
Gift a Flow
The killer onboarding mechanism. Gift a Flow lets any user create a pre-configured flow URL that, when opened by a new user, sets up a working Slipstream workspace with zero friction—no sign-up required for the preview. “Here’s a link to see the weather and garden conditions at our community garden.” The recipient sees live data immediately and can claim the flow with a single sign-up step.
Interest Calibrator
Progressively refines the user’s interest profile through both explicit signals (preference surveys, rill activation) and implicit signals (which rills they spend time on, which alerts they act on, which digests they read). Interest Calibrator avoids the “cold start” problem by starting with broad interest categories and narrowing over time, never asking more than one question per session.
Migration Assistant
Helps users bring their data from other platforms into Slipstream. Migration Assistant imports historical weather data from Weather Underground, garden logs from Gardenate, health data from Apple Health, bird lists from eBird, and observations from iNaturalist. It maps external data schemas to Slipstream’s internal format, preserving history so users don’t start from zero.
Achievement System
Milestones and discovery badges that encourage exploration without gamifying the experience into triviality. Achievement System awards badges for meaningful actions—“First Composition” (composed two rills), “Night Observer” (used nocturne mode at 2 AM), “Full Season” (used Slipstream through an entire growing season). Badges are private by default; share only if you want to.
agenda (things to be done)
Agenda is where observation becomes action. While other rills watch the world, agenda manages what you’re going to do about it—rill-triggered tasks (“soil dry, water garden”), seasonal project planning, reusable checklists, condition-based reminders, multi-step project tracking, and periodic reviews with data-backed retrospectives. It bridges the gap between knowing and doing.
terr-03 → kr-02 → kr-04 → ad-01 → ts-01
Soil moisture rill detects low reading, threshold engine triggers, event cascade fires, triggered task creates “Water garden today,” and calendar sync adds it to Google Calendar.
| ID | Name | One-liner |
|---|---|---|
| ad-01 | Triggered Tasks | Auto-generated tasks from rill conditions: “Soil dry → Water garden” |
| ad-02 | Seasonal Planner | Long-arc project planning tied to natural calendar |
| ad-03 | Checklist Templates | Reusable checklists: hive inspection, garden walk, foraging prep |
| ad-04 | Reminder Engine | Timed and condition-based reminders from any rill trigger |
| ad-05 | Project Tracker | Multi-step projects with rill-informed milestones and dependencies |
| ad-06 | Review & Reflect | Weekly/monthly/seasonal review with data-backed retrospective |
| ad-07 | Kanban Board | Visual task board with customizable columns and swimlanes |
| ad-08 | Recurring Tasks | RRULE-based recurring tasks with flexible scheduling |
| ad-09 | Time Blocking | Calendar-integrated focus blocks for deep work and rill review |
| ad-10 | Habit Tracker | Streak counting, consistency scoring, and habit-rill correlation |
Triggered Tasks
Automatically creates tasks from rill conditions—the bridge between observation and action. When soil moisture drops below threshold, a “Water garden” task appears. When the frost forecast triggers, “Cover tender plants” gets added. When hive weight drops suddenly, “Check for swarm activity” lands in your inbox. Each triggered task includes context from the rill that generated it: the reading, the trend, the recommended response.
Seasonal Planner
Long-arc project planning tied to the natural calendar rather than arbitrary fiscal quarters. Seasonal Planner helps organize multi-month efforts—garden season prep (February–April), honey harvest (July–September), winter garden cleanup (October–November)—with milestones that shift based on actual phenology data rather than fixed dates. Spring comes early this year? The planner adjusts.
Checklist Templates
Reusable checklists for recurring activities—hive inspection (check queen, count frames, assess stores), garden walk (soil moisture, pest check, harvest ready), foraging prep (gear, map, weather check, safety plan). Templates are community-shareable and can include rill data auto-fills: the hive inspection checklist pre-populates today’s weight reading and weather conditions.
Reminder Engine
Flexible reminders that can be time-based (“remind me at 6 PM”) or condition-based (“remind me when soil moisture recovers above 40%”). Reminder Engine integrates with kairos for condition monitoring and nuntius for delivery, supporting snooze, recurrence, and escalation. A reminder isn’t just a ping—it includes the current rill context so you know why you’re being reminded.
Project Tracker
Multi-step projects with rill-informed milestones and dependency tracking. Project Tracker manages complex efforts like “Build raised beds” (materials → construction → soil fill → planting) with milestones that can be tied to rill conditions: “Planting milestone unlocks when soil temperature reaches 60°F.” Progress visualization shows completed, in-progress, and blocked steps.
Review & Reflect
Periodic review sessions backed by actual data. Review & Reflect generates weekly, monthly, and seasonal retrospectives that pull from task completion rates, rill trends, health metrics, and journal entries. “This week you completed 8 of 12 garden tasks, soil moisture averaged 42%, and you logged 3 species observations. Last week: 5 tasks, 38% moisture, 1 observation.” Reflection with receipts.
Kanban Board
Visual task management with customizable columns (To Do, In Progress, Blocked, Done) and swimlanes by project or category. Kanban Board renders tasks from all agenda rills into a drag-and-drop board that supports filtering by workspace, priority, and rill source. The visual layout reveals bottlenecks: too many items in “Blocked” means something upstream needs attention.
Recurring Tasks
RRULE-based recurring tasks for things that happen on a schedule—weekly hive inspections, biweekly lawn mowing, monthly garden journal entries, seasonal equipment maintenance. Recurring Tasks handles complex patterns (every other Tuesday, first Monday of the month, every 10 days) and generates task instances ahead of time so they appear in your agenda before they’re due.
Time Blocking
Calendar-integrated focus blocks that reserve time for deep work, rill review, and outdoor activities. Time Blocking syncs with ts-01 (Calendar Sync) to place blocks on your calendar and integrates with nu-05 (Quiet Intelligence) to suppress notifications during focus periods. It suggests optimal timing based on your circadian rhythm from sa-01 and weather conditions from aer-01.
Habit Tracker
Tracks daily habits with streak counting, consistency scoring, and—uniquely—correlation with rill data. Habit Tracker answers questions like “Do I garden more on days when the weather is clear?” and “Does my sleep quality correlate with whether I did evening stretching?” It turns habit tracking from simple checkboxes into a data-informed understanding of what drives your behavior.
action (action, doing, performance)
Action is the most cross-cutting family in Slipstream. Its rills observe the entire system—reading from every domain via a read-only registry—and synthesize what they see into daily briefings, anomaly narrations, decision advice, priority rankings, and emergency coordination. Where fo-01 (Narrator) turns one rill’s output into prose, action turns the entire system’s state into narrative intelligence.
[all-rills] → ac-01 → ac-03 → nu-01 → nu-04
AI narrator observes all rill states, decision advisor extracts actionable insights, digest composer assembles the briefing, voice briefing speaks it aloud.
| ID | Name | One-liner |
|---|---|---|
| ac-01 | AI Narrator | Autonomous observation across all rills, synthesized narrative |
| ac-02 | Anomaly Narrator | Prose explanation of detected anomalies with context and history |
| ac-03 | Decision Advisor | Data-backed recommendations for time-sensitive choices |
| ac-04 | Priority Ranker | Daily prioritization of tasks and attention based on rill signals |
| ac-05 | Emergency Coordinator | Multi-rill synthesis during critical events (fire, flood, storm) |
| ac-06 | Pattern Storyteller | Long-term pattern narration: “Your garden this year vs. last” |
| ac-07 | Correlation Finder | Cross-domain signal correlation that humans might miss |
| ac-08 | Scenario Modeler | “What if” analysis: model outcomes of different decisions |
| ac-09 | Attention Manager | What deserves your attention right now? Ranked signal digest |
| ac-10 | Learning Journal | System self-improvement: what predictions were right, what missed |
AI Narrator
The most cross-cutting rill in the entire system. AI Narrator reads from all families via a read-only registry, synthesizes what it observes, and produces a coherent narrative about the state of your world. Not just data summaries—genuine narrative intelligence: “The soil dried out faster than expected this week because temperatures ran 8 degrees above average while you were traveling. Combined with the hive weight plateau, this suggests the nectar flow is winding down earlier than last year.” It connects dots that individual rills can’t.
Anomaly Narrator
When kr-01 detects an anomaly, Anomaly Narrator explains it in plain language with context. Not just “soil moisture anomaly detected” but “Soil moisture dropped 15% overnight—unusual for this time of year. The most likely cause is yesterday’s 12 mph winds combined with above-average temperatures. Similar drops last August preceded a 3-day recovery cycle.” It transforms a statistical flag into an understandable story.
Decision Advisor
Surfaces data-backed recommendations for time-sensitive decisions. Should you water the garden today or wait for tomorrow’s rain? Is this a good day for outdoor work? Should you harvest honey now or wait another week? Decision Advisor synthesizes weather forecasts, soil data, phenology, and task schedules into clear recommendations with confidence levels and reasoning. Signals, not commands—always the user’s call.
Priority Ranker
Ranks today’s tasks and attention items by urgency, importance, and opportunity cost. Priority Ranker considers weather windows (“outdoor work today, desk work tomorrow since rain is coming”), biological timing (“the swarm cells need attention now”), and deadline proximity. It produces a ranked daily priority list that adapts as conditions change throughout the day.
Emergency Coordinator
Multi-rill synthesis during critical events—wildfire smoke, flooding, severe storms, power outages. Emergency Coordinator pulls from weather, air quality, road conditions, utility status, and community reports to produce a unified situation picture with actionable guidance. It overrides normal notification settings to ensure critical information reaches the user through every available channel.
Pattern Storyteller
Narrates long-term patterns that are invisible in daily data. Pattern Storyteller compares seasons, years, and multi-year trends: “Your garden produced 30% more this summer than last, likely because you started 2 weeks earlier and the spring was unusually mild. Bee activity peaked 10 days later than 2024, correlating with the delayed blackberry bloom.” It turns data archives into stories of change.
Correlation Finder
Discovers cross-domain signal correlations that humans might miss. Correlation Finder continuously analyzes relationships between rill outputs: “Your sleep quality correlates strongly with barometric pressure changes” or “Hive weight gains are 40% higher when wind speed is below 5 mph at dawn.” It surfaces these findings as insights, not conclusions—correlation isn’t causation, and the narration says so.
Scenario Modeler
“What if” analysis that models outcomes of different decisions. Should you plant tomatoes this weekend or wait two more weeks? Scenario Modeler runs both timelines against weather forecasts, soil temperature trends, and historical frost dates to show likely outcomes. It doesn’t predict the future—it shows the range of possibilities so you can make an informed choice.
Attention Manager
A curated signal digest that answers: “What deserves my attention right now?” Attention Manager ranks all active signals across all domains by relevance, urgency, and novelty, then presents the top items as a focused list. It filters out noise, groups related items, and ensures you see the one thing that matters most before the ten things that can wait. Attention is finite; this rill respects that.
Learning Journal
System self-improvement through prediction tracking. Learning Journal records what action rills predicted (frost probability, harvest timing, anomaly cause) and later checks what actually happened. Over time, it builds a calibration record: which predictions are reliable, which are consistently off, and where the models need improvement. It’s how Slipstream gets smarter through honest self-assessment.
tempus (time, season, occasion)
Tempus is the temporal multiplier—it transforms any content rill’s output from a snapshot into a story told through time. Feed it any metric and it reveals the trend, compares it to last season, flags anomalies against historical baselines, forecasts what comes next, reconstructs past conditions, and detects changes between two points in time. Tempus doesn’t own any data domain; it owns the dimension of time itself.
terr-03 → tm-01 → tm-02 → tm-04 → ac-06
Soil moisture feeds the trend analyzer, seasonal comparator shows this year vs. last, prediction engine forecasts next week, pattern storyteller narrates the full seasonal arc.
| ID | Name | One-liner |
|---|---|---|
| tm-01 | Trend Analyzer | Direction, rate of change, and inflection points for any metric |
| tm-02 | Seasonal Comparator | This season vs. last year vs. 10-year average for any metric |
| tm-03 | Anomaly Detector | Flag readings significantly outside historical range |
| tm-04 | Prediction Engine | Forecasting from historical patterns—when will X happen next? |
| tm-05 | Historical Reconstruction | What were conditions at this location on a past date? |
| tm-06 | Change Detection | Satellite/data comparison between two dates—what changed? |
Trend Analyzer
Takes any rill’s time-series data and extracts the signal: is it going up, down, or sideways? How fast? Where are the inflection points? Trend Analyzer applies moving averages, linear regression, and change-point detection to reveal the trajectory hidden in noisy readings. Soil moisture isn’t just 42% right now—it’s been declining at 3% per day since Monday, and at this rate hits your alert threshold Thursday morning.
Seasonal Comparator
Puts current conditions in seasonal context by comparing them to the same period last year, five years ago, or the long-term average. Seasonal Comparator answers: “Is this spring warmer than usual? Is the bloom earlier? Is rainfall on track?” It overlays time series from different years on the same axis, making it immediately visible whether this season is normal, early, late, or unprecedented.
Anomaly Detector
Flags readings that are significantly outside their historical range for this time of year and location. Unlike kr-01 (which detects anomalies against recent baselines), tempus’s Anomaly Detector uses deep historical context: 70°F in February in Portland is unremarkable for kr-01 but historically extraordinary for tm-03, which knows the 30-year February average is 47°F. Historical depth reveals what recent data cannot.
Prediction Engine
Forecasts future values from historical patterns. When will the first frost likely occur? When will soil temperature reach planting threshold? When will the blackberries be ready to pick? Prediction Engine uses seasonal decomposition, regression models, and pattern matching against historical data to answer “when will X happen next?”—always with confidence intervals, never false certainty.
Historical Reconstruction
Answers: “What were conditions at this location on [past date]?” Historical Reconstruction queries archived data from weather, satellite, phenology, and other sources to reconstruct the state of a place at any point in the past. Useful for context (“What was the weather on my wedding day?”), analysis (“What were soil conditions when I planted last year?”), and storytelling (“What did this landscape look like 10 years ago?”).
Change Detection
Compares data between two points in time to reveal what changed. Change Detection works with satellite imagery (NDVI changes showing deforestation, development, or seasonal shifts), data snapshots (population changes, species lists, land use), and rill outputs (how has your garden’s canopy coverage changed since spring?). It produces before/after comparisons with quantified differences and highlighted change areas.