Everything your ground is already telling you
AgriVerse resolves satellites, sensors, and weather down to the individual field, then keeps the record of what you did about it.
Eight capabilities, one field record
Each one is useful on its own. Together they answer the only question that matters on a Tuesday morning: which field needs me today?
Crop health from space
Every mapped field, every clear pass, without buying a single sensor.
Sentinel-2 crosses your farm roughly every five days at 10 m resolution. We pull the bands, mask the cloud, clip to your boundary, and compute the vegetation indices per field, then keep the whole history so you are looking at a trend, not a snapshot. Landsat 9 fills in when Sentinel is clouded out.
What you get
- NDVI, EVI, NDRE, NDWI and SAVI per field
- Mean, min, max and standard deviation across the boundary
- Year-over-year comparison against the same window last season
- History backfilled from before you signed up
What it won't do
Optical imagery cannot see through cloud. A clouded pass is marked as such rather than interpolated into a number you might act on.
Boundaries that mean something
Draw it once, and every screen agrees on how big the field is.
Trace a boundary on the map in about a minute, or import what you already have: KML, shapefile, GeoJSON, and the exports most machinery terminals produce. Area is computed from the geometry, not typed in, so the ledger, the bill, and the yield figures all reconcile.
What you get
- Import from KML, shapefile, GeoJSON and ISOXML
- Area, centroid and bounds computed from the polygon
- Soil profile, irrigation type and crop attach to the field
- Boundary edits are versioned, never silently overwritten
What it won't do
Imported geometry is only as good as its source. We show the computed area next to any area you were carrying, so a mismatch surfaces immediately.
Management zones
Treat the part of the field that will answer, not the average of it.
The index raster is split into productivity zones by quantile, so a 30-hectare block resolves into the four sub-blocks that actually behave differently. Export the zones as a prescription, or use them to decide where to walk first.
What you get
- Four productivity zones per field, ranked weakest to strongest
- Zone area and mean index, so the split is defensible
- Prescription export for variable-rate application
- Zones recomputed as the season develops, not fixed in March
What it won't do
Zones describe where the canopy differs, not why. Ground-truth the weakest zone before you commit a variable-rate pass to it.
Irrigation you can defend
Rainfall in, evapotranspiration out, and the deficit you're actually carrying.
Soil moisture from your probes, rainfall from the station or the forecast, and reference ET on one water balance per field. When a field approaches its refill point you get an alert with an estimated requirement, and once you irrigate, the volume goes into the record.
What you get
- Per-field water balance with refill-point estimate
- Applied volumes recorded from flow meters or logged by hand
- Applied against crop requirement, summarised per season
- Over-application flagged, not just under-application
What it won't do
Refill points come from a soil-texture model unless you supply lab results. Real plant-available water varies; treat the estimate as a prompt to check, not a set point.
Sensors without the babysitting
Probes, stations and meters that tell you when they've stopped talking.
Register a device, point it at the ingest endpoint over MQTT, HTTPS or LoRaWAN, and its telemetry lands against the field it sits in. Thresholds turn a reading into an alert, and a missed heartbeat window is itself an alert, because a silent probe is the most expensive kind.
What you get
- MQTT, HTTPS and LoRaWAN ingest
- Soil moisture, temperature, EC, rainfall, wind and solar
- Per-metric thresholds with severity
- Battery, signal and heartbeat monitoring per device
What it won't do
We are sensor-agnostic on purpose, which means we cannot vouch for someone else's calibration. Compare a new probe against a hand sample in its first fortnight.
An append-only field record
The history you show a buyer is the history you were working from.
Tillage, planting, irrigation, nutrition, crop protection, scouting, harvest: each recorded against a field with who, when, how much, and on what area. Records are append-only: a correction creates a new linked entry and the original stays visible, struck through, pointing at its replacement.
What you get
- Every operation attributable to a person and a timestamp
- Products, rates, equipment and cost captured per action
- Corrections link forward; nothing is ever edited or deleted
- Exportable for certification and buyer audits
What it won't do
An append-only log proves what was recorded and when, not that what was recorded is true. It removes the argument about tampering, not the need for honest entry.
Weather that speaks agronomy
Not a forecast widget. The numbers a spray or irrigation decision turns on.
Per-farm forecast and history with growing-degree-days accumulated against the crop actually in the ground, reference evapotranspiration for the water balance, and advisories when a heat run, a frost, or a wet week is going to collide with a growth stage that matters.
What you get
- 7-day forecast and 30-day history per farm
- Growing-degree-days against the planted crop's base temperature
- Reference ET for irrigation scheduling
- Heat, frost and heavy-rain advisories tied to growth stage
What it won't do
Forecasts are gridded model output, not an on-farm observation. Where you have a station, we show its reading next to the forecast rather than instead of it.
Season reporting
What the season cost, what it yielded, and where the water went.
Modelled yield per field and per crop, input cost broken down by category, water applied against requirement, and the operations calendar month by month. The numbers come from the ledger, so the report is a view of the record rather than a second set of books.
What you get
- Yield modelled through the season, actual after harvest
- Input cost by category, in your own currency
- Water applied against crop requirement, per crop
- Operations per month, with area covered
What it won't do
Yield modelling is a projection from canopy development and weather. It sharpens through the season and is never a substitute for a weighbridge.
See it on your own fields.
Draw your boundaries and the satellite record is waiting, including the passes from before you signed up. 21 days free, no card, and every field you map is exportable the moment you ask.