How and where it can be used?
The everyday farming jobs the score is used for.
Fifteen everyday farming decisions this score is used for.
Buying or leasing new acreage
Before signing, check whether the plot actually holds moisture and grows well, instead of trusting the seller's photos.
Approving a farm loan
A lender gets an instant, defensible read on the land pledged as collateral without sending anyone to the village.
Pricing a crop insurance policy
Set the premium on a parcel at quote time, so good land is not charged the same as marginal land.
Settling a drought claim
See whether the field really dried out, and when, before paying or rejecting a claim.
What's new?
The legacy pre-construction workflow, and what Rootfifteen actually replaces.
How this was done before — and what changes.
A live read while you are still in the meeting.
No site visit, no imagery contract.
Every score breaks down into measured values.
How it works?
You send a coordinate. We read live satellite and climate data. You get one score back.
Five steps, every request.
- 01Inputlat / lng + mode: farming
- 02ObserveLive pixel + climate sampling
- 03DeriveIndices and stress signals
- 04FuseWeighted, deterministic model
- 05Output0–100 score + band + components
Built on open Sentinel-2, Copernicus DEM, NASA POWER and USGS data — read live at your coordinate, never copied from a nearby area.
Four open data streams, one score.
10–20 m multispectral imagery for vegetation, water and surface indices.
~90 m global elevation for slope, relief and terrain shape.
Temperature, rainfall and solar regime at the coordinate, not a regional average.
Recorded seismic events within 200 km over the last five years.
Three fields in. That is the whole input.
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"farming"Tells the model which decision you are making. Farming mode weights surface vigour and climate most heavily.
POST /api/public/v1/land-intel/assess
{
"latitude": 12.9716,
"longitude": 77.5946,
"mode": "farming"
}200 OK
{
"status": "ok",
"engine": "land-intel.v1",
"mode": "farming",
"rootfifteenScore": 74.2,
"classification": "Healthy",
"components": {
"surface": 78,
"climate": 71,
"terrain": 82,
"georisk": 90
},
"observations": {
"ndvi": 0.62, "ndmi": 0.21,
"ndwi": -0.18, "ndbi": -0.09,
"elevationM": 912, "slopePercent": 1.8,
"meanTemperatureC": 24.1,
"annualPrecipitationMm": 921
},
"verticalMetrics": [
{ "key": "canopy vigour",
"label": "Canopy vigour",
"value": 25.4, "display": "25.4 / 100",
"band": "Bare or degraded ground" }
],
"generatedAt": "2026-07-11T12:04:22Z"
}What you get back.
Exactly what the live demo report shows for a coordinate:
- 1The Rootfifteen score, 0–100, for the exact coordinate.
- 2The classification band that goes with it, from Optimal to Degraded.
- 3Four component sub-scores: surface, climate, terrain and georisk.
- 4Farming indicators: canopy vigour, root-zone moisture, irrigation dependence, thermal season, waterlogging risk, topsoil erosion risk, mechanisation suitability and sealed surface.
- 5Every raw measured value behind those indicators, exactly as the API returns it.
- 6A written summary of what the numbers mean for the plot.
- 7A PDF of the whole report.
What the number means.
Every response carries a classification alongside the score. The score ranges from 0 to 100; these are the exact bands the API returns for farming.
Now 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.