We find the fields
Every parcel drawn straight from satellite — automatically, even where no cadastre exists. Without this map, nothing downstream has anything to attach to.
Live: 69M parcels — including 1.6M in Ukraine, drawn from scratch.
The chain
Find the field. Read the crop. Call the tonnes.
Each link only works if the one before it does — which is why most providers stop at the first. We run the whole chain, and we run it while the crop is still in the ground.
We read the crop — early
Wheat, maize, rapeseed, sugar beet, barley, rye, flax and more, field by field. Most maps land after harvest; ours read the crop while it is still growing.
17+ classes, mid-season — months before the print.
We call the tonnes Pilot
Known parcel, known crop, watched all season — the ingredients yield needs. We build yield on top of that chain, scoped to your geography and crop.
Pilots running — on request for your region.
Why it travels
Foundation models that read farmland across borders.
We read a full season of satellite signal through geospatial foundation models. That's why a model calibrated on one country still reads the next — where field-data-hungry approaches collapse the moment they cross a border.
Observe
A whole season, every field
Continuous coverage across the country — not a single snapshot.
Delineate
Boundaries from pixels
Parcel geometry inferred directly — no cadastre required.
Classify
Foundation embeddings → crop type
Cross-domain by design; the domain gap closes instead of breaking.
Estimate
Crop + season → tonnes
Yield built on the parcel and crop already established upstream. Pilot.
Deliver
Mid-season, per field
API, tiles or bulk — grouped to your taxonomy.
one model · France → Spain → Portugal, no retraining
Who uses it
One field-level data layer, priced to your decision.
Commodity traders
Early, wall-to-wall production signal — acreage and crop mix before official stats print.
Crop & parametric insurers
Field-level exposure, planting verification and faster, evidenced claims.
Ag lenders & banks
Collateral you can see — what's planted on the parcels behind every loan.
Seed & input companies
Demand mapped to real crop rotations, region by region.
Agtech & food supply
A ready field-level base layer to build products and traceability on.
Governments & statistics
Independent, timely crop area estimates — including where declarations are thin.
Real output
This is what lands in your hands.
Not mock-ups — live tiles from our production maps, exactly as they render today.
Czech Republic → Poland — the map doesn't stop at the border
One model, both sides of the line. Fields stay typed across the frontier, where border-hopping approaches break.
Spain — Extremadura, wall to wall
Mérida and the Guadiana valley. Irrigated row crops, cereals and permanent crops, each parcel typed.
Ukraine — 1.6M parcels, drawn from scratch
Vinnytsia region. Boundaries detected automatically — no cadastre, no field visits.
France — the Beauce, Europe's grain belt
South of Étampes. Wheat, rapeseed and barley across thousands of parcels in a single pass.
Poland — Wielkopolska, plot by plot
East of Poznań. Narrow strip farming, sugar beet and cereals — resolved down to the individual plot.
Coverage
Live today, and expanding by the season.
Interactive per-field maps you can pan right now — crops and boundaries, updated as new imagery lands.
Europe · 25+ countries
Ukraine · 1.6M fields
Kazakhstan · 1.9M fields
Luxembourg · 2023–2025
Poland · Germany · France · Spain
Netherlands · Baltics · Balkans
+ your region on request
FAQ
Questions desks ask us first.
How is this different from official crop statistics?
Official numbers arrive late, aggregated to a region, and only after the season closes. We deliver the same picture per field and mid-season — months earlier, at a resolution you can act on, in countries where no reliable statistics exist at all.
Which countries do you cover?
30+ countries across Europe, plus Ukraine, Kazakhstan and more — 163M hectares live today. If your area of interest isn't on the map yet, we add regions on request; a new country is typically a matter of weeks, not a new product.
How accurate is it?
Our maps are produced from satellite and validated against ground truth. Accuracy varies by crop, region and how early in the season you ask — winter crops read earlier and cleaner than spring crops. We're happy to share validation figures for your geography and crops before you commit.
Do you need my field boundaries or farm data?
No. We detect field boundaries directly from satellite — that's the first link in the chain — so we can start anywhere, with no cadastre and no field visits. If you already have your own parcels, we can classify crops on those instead.
How do you deliver the data?
However fits your stack: a REST API, vector tiles for your map, or bulk export (GeoJSON, GeoPackage, Parquet). Crops come in a raw taxonomy and can be grouped to your own classes on delivery.
What about yield?
Yield is the third link in the chain and is currently in pilot. Because we already know the parcel and the crop and watch it all season, we build yield on top of that — scoped to a specific geography and crop. Tell us your region and we'll scope a pilot.
How does pricing work?
Pricing depends on the geography, how often you need updates, and the delivery format. Request access with your area of interest and we'll come back with a scoped quote — no obligation.
Request access
Put a satellite over your acreage.
Tell us the geography and the decision. We'll send back a sample map for your region and a short call.
✓ A field-level sample for your area of interest
✓ Coverage, cadence and format that fit your stack
✓ No obligation — just the data, on your fields
✓ Request received.
Thanks — we'll pull a sample map for your region and get back to you from hello@cropmap.ai, usually within a day.