How it works
There's a real pipeline under the hood — satellite, radar, weather and soil, run through agronomy models and calibrated against reality, nothing hidden. Here's the whole path, from a boundary you draw to the call you make.
No sensors to install, no data to import, no key to request.
Trace a boundary on the map, drop a pin and let Row Wise detect the field edges, or import a shapefile / KML. That's the entire setup.
Sentinel‑2, NASA HLS and Sentinel‑1 radar auto‑import and clip to your boundary — cloud‑screened, no account and no API key.
Open the app to a ranked, dollarized to‑do list. Tap done, log an irrigation, drop a scouting pin — and it learns your season.
The inputs
Every source below is public and documented. That's why there's no imagery to license — and why you can trace any number back to where it came from.
10 m multispectral imagery every few days — the backbone of NDVI, NDRE, MSAVI & CIre vigor.
C‑band SAR that sees through cloud — canopy structure and surface moisture in the weeks optical can't.
Harmonized Landsat‑Sentinel fills gaps and adds haze‑proof passes, with a CMR fallback when needed.
Soil survey available‑water capacity — the bucket size behind the irrigation checkbook.
Global soil fertility & carbon context where survey detail runs out.
Hourly weather & ET₀, sub‑seasonal ensembles and a decade of history — no key, portable by design.
High‑resolution aerial basemap imagery for context and boundary drawing.
The Cropland Data Layer masks your crop for honest county‑level benchmarking.
Weather is sourced through one adapter — self‑host or commercial, without touching a single formula
Documented, not magic
Where a number comes from a formula, it's a documented one — FAO‑56 for the water balance, FAO‑33 for yield response, AgDRIFT for spray drift, DD60 heat units for growth stage. Where it comes from a learned model, the model is conventional and trained on a held‑out year so you can see how well it actually did.
FAO‑56 · FAO‑33 · AgDRIFT · GDD/DD60 — cited, not hidden.
Gradient‑boosting & random‑forest on engineered features — weather‑driven and leak‑free.
Fiber quality ties to gin HVI; yield ties to combine maps; maturity ties to scout counts.
We show our work
The NDVI forecast is validated against a year the model never trained on — the dashed prediction against the solid truth. When a model is uncertain, the app widens the band and says so.
The principles
Every number traces to a public source and a named method. If you want to know how the health score got to 73, the app will tell you.
Estimates come as ranges. Thin clouds, sparse observations and RGB‑only imagery are flagged — never quietly averaged into a false certainty.
Fields are owner‑scoped and access is fail‑closed. Your boundaries and records are yours.
Public data and one weather adapter mean the whole thing can be self‑hosted — no vendor lock on the pixels.
Draw a field and watch public imagery and weather turn into your first ranked action feed.
Free to join the waitlist · no spam · we will email you the day it opens