A coordinate on a field.A crop-viability answer.
For insurers, lenders, agri-input companies and food supply teams: replace spot-checks, satellite-tile pipelines and subjective field reports with one objective 0–100 score per coordinate — computed from live Sentinel-2 imagery, terrain and climate reanalysis.
What the API does for agriculture.
No shapefile, no polygon, no farm registry. A latitude/longitude pair is enough to get a viability read on a parcel you have never visited.
NDMI moisture reserve plus climate reanalysis expose dry-down and heat-degree build-up before it shows up in yield.
NDVI is sampled on the actual Sentinel-2 pixel at 10 m, with cloud and shadow pixels rejected rather than smoothed over.
Because the model is deterministic, thousands of parcels can be ranked on the same axis and re-scored on the same axis next season.
Coordinate to decision, step by step.
Every agriculture request runs through the same five stages. Nothing is cached from a neighbouring tile and nothing is invented — if a pixel is unusable, the response says so.
- 01Inputlat / lng + mode: agriculture
- Single HTTP POST
- No shapefile or polygon
- No upload, no GIS stack
- 02ObserveLive pixel + climate sampling
- Sentinel-2 red / NIR / SWIR at 10–20 m
- SCL cloud & shadow gate
- Climate reanalysis at the point
- DEM five-point stencil
- 03DeriveIndices and stress signals
- NDVI vigour · NDMI moisture
- NDWI standing water penalty
- NDBI sealed-surface penalty
- Heat-degree and drought build-up
- 04FuseWeighted, deterministic model
- Surface weighted highest
- Climate second
- Terrain and georisk as modifiers
- No random seeds
- 05Output0–100 score + band + components
- Machine-readable JSON
- Component sub-scores exposed
- Same coordinate → same answer
No random seeds. No hidden drift. If you got a 74 last Tuesday, you'll get a 74 next Tuesday — unless the ground itself has changed. That is what makes the score defensible in an underwriting memo, a planning submission, or a board pack.
Three fields in. That is the whole input.
No shapefiles. No polygons. No uploads. No GIS stack to maintain.
latnumberLatitude in decimal degrees, −90 to 90. The exact point of the field, plot centroid or sample location.
lngnumberLongitude in decimal degrees, −180 to 180.
mode"agriculture"Tells the model which decision you are making. Agriculture mode weights surface vigour and climate most heavily.
POST /v1/land-intel/assess
{
"lat": 12.9716,
"lng": 77.5946,
"mode": "agriculture"
}200 OK
{
"score": 74,
"classification": "HEALTHY",
"band": "Productive",
"mode": "agriculture",
"components": {
"surface": 78,
"climate": 71,
"terrain": 82,
"georisk": 90
},
"observations": {
"ndvi": 0.62, "ndmi": 0.21,
"ndwi": -0.18, "ndbi": -0.09,
"elevation_m": 912, "slope_pct": 1.8
},
"generated_at": "2026-07-11T12:04:22Z"
}Enterprise-grade data. Open licences. Zero cost.
Every agriculture score is built on the same public Earth-observation feeds that power commercial GIS stacks sold for tens of lakhs to crores per annum. We expose them directly through one simple HTTP call.
10–20 m multispectral surface reflectance from ESA's Copernicus Sentinel-2 constellation, served as cloud-optimised GeoTIFFs via the AWS Earth Search STAC catalogue. Every request samples the actual pixel at your coordinate for red, green, NIR and SWIR bands.
GLO-90 digital elevation model from the Copernicus DEM programme, sampled through a five-point stencil so slope and local relief are measured, not assumed. The centre point plus four surrounding points gives a true local gradient.
Every sampled pixel is checked against the Sentinel-2 SCL. Cloud, cloud shadow, thin cirrus, saturated and no-data pixels are rejected outright — not smoothed over — so the score only sees ground truth.
Open-Meteo historical climate reanalysis provides temperature, precipitation and humidity baselines at the coordinate. The 5-year seismic catalogue supplies catalogued ground-motion frequency and peak magnitude within 200 km.
No random seeds. No hidden drift. If you got a 74 last Tuesday, you'll get a 74 next Tuesday — unless the ground itself has changed. That's what makes the score defensible in an underwriting memo, a planning submission, or a board pack.
What you get back.
A plain-English report anyone can read. No remote-sensing degree, no GIS software, no manual report.
One simple number that tells you how suitable this point is for crops.
What the ground actually looks like right now — green cover, moisture, and surface type.
A simple label like Healthy, Productive, or Stressed, plus a recommended action.
See the four drivers behind the score — vegetation, climate, terrain, and georisk.
The result comes as clean JSON that any software can read.
The parameters behind the score.
Every response is one score plus the measured parameters that produced it. Which parameters come back depends on the vertical you query.
How much live, photosynthesising canopy is on the ground.
Water held in the canopy and topsoil.
Open or ponded water on the parcel.
How much of the parcel is roads, roofs or bare impervious surface.
Temperature envelope and precipitation over recent seasons.
Slope and relief across the parcel.
What those outputs do in real life.
Re-score affected coordinates after a weather event. When the score crosses the agreed band, the payout fires — no site visit, no dispute over what the field looked like.
A loan officer enters a plot location and gets a viability band on screen. Strong land clears instantly; marginal land is referred with the exact component that caused it.
Score supplier farms quarterly and surface degrading or converted land before it becomes a compliance failure in a sourcing audit.
Rank dealer-network parcels by potential so premium seed, fertiliser and agronomy hours go where they can actually pay back.
Pick a coordinate. See the real output.
These are live calls to the production agriculture engine — real Sentinel-2 imagery, real elevation, real climate and seismic records. Nothing here is mocked.
Intensively farmed alluvial plain
How green and actively growing the crop stand is
Water held in the canopy and topsoil
Ponding, flooding or open water on the parcel
Roads, roofs and bare impervious surface eating arable area
Water the parcel receives in a year
Growing-season heat load
Steepness — affects erosion and machinery
When the satellite last saw this parcel
Fused live from Sentinel-2 L2A optical imagery, Copernicus GLO-90 elevation, NASA POWER climate reanalysis and the USGS seismic catalogue — returned in a single JSON response, typically in about a second.
How this was done before — and what changes.
Every row is a task agriculture teams already do today. The left column is the traditional method. The right column is the same task with one API call.
Ten bands. One axis. Zero ambiguity.
The 0–100 axis is partitioned into ten calibrated bands, each mapped to a defensible agriculture decision — identical across every coordinate on Earth.
The operational workflows Rootfifteen replaces.
Parametric insurance payouts
Re-score affected coordinates after weather events and trigger payouts when parcels drop below agreed bands.
Farm credit underwriting
Auto-approve high-quality parcels, flag marginal applications, and decline weak land with a data-backed explanation.
Supply chain land verification
Score supplier farms quarterly and surface degraded or protected-land risk before it becomes a compliance problem.
Precision input targeting
Rank dealer-network parcels by potential so premium products and advisory time go where they can actually perform.
Drop the API into the systems that already run agriculture decisions.
Rootfifteen is one HTTP endpoint. That means it fits inside any software that already touches agriculture workflows — no new dashboard, no new map stack, no internal GIS team required.
Bind policies, score parcels at quote time, and trigger parametric payouts after weather events without loss adjusters.
Auto-approve strong land, flag marginal applications, and explain declines with component sub-scores.
Rank parcels by vigour and moisture so agronomists and input dealers spend time where it pays back.
Score supplier farms quarterly and catch degraded or converted land before audits do.
Monitor crop health across districts and prioritise drought relief, irrigation, and extension spend.
Build a reproducible, auditable land-quality factor into treaty pricing and risk selection.
Every agriculture call returns a 0–100 score, a calibrated band, a machine-readable classification, and the component sub-scores that explain it. Wire it into your underwriting engine, your site-screening workflow, or your command map exactly like any other REST API.
Same coordinate. Different decision.
For developers, real-estate underwriters, infrastructure planners and property insurers: compress terrain, hydrology and ground-motion risk …
Read the construction pageFor defence planners, border and disaster-response agencies and logistics staff: turn a coordinate into an operability score that captures h…
Read the defence pageNow try it in your own stack.
Tell us what you are building and we will send you endpoint docs, a key, and limits that fit your use case.