Cropiris turns an ordinary phone photograph into a root-zone water balance, a disease identification and a date to act on — for growers who want a figure they can defend, not a colour-coded impression.
No card required. Irrigation advice and measurements stay free, for good.
Irrigate within two days. 34 mm of readily available water left, and the forecast adds none.
What it does
Five stages run over every capture, and the last one — the part that writes the sentence you read — is never allowed to originate a diagnosis of its own.
A FAO-56 soil water balance runs every day against your field's own weather: reference evapotranspiration, rainfall, and the depletion fraction for the crop. You get how much water is gone and how long until it matters — not a wetness score.
Leaf symptoms are classified by a specialist model, not by a chatbot. When it is unsure the case goes to an agronomist, and the correction is kept beside the machine answer rather than quietly replacing it.
Canopy cover measured from your photograph sets the crop coefficient, so the water balance tracks the crop you actually have — not a textbook curve for the crop you planted. This is the reason a camera belongs in an irrigation tool.
Captures queue on the device and upload when you are back in range. Nothing is lost because a shed has thick walls, and a capture taken offline still carries the tilt, exposure and timing that make it comparable.
Grant an agronomist access to one farm without handing over the rest. Their captures belong to the farm they document, so when the contract ends the record stays with the grower.
Each figure names the measurements behind it, the model version that produced it, and whether a person has reviewed it. Nothing is presented as certain because it was expensive to compute.
A season, measured
Every day the balance subtracts what the crop used and adds what fell. The line is where your soil actually is; the dashed part is where the forecast says it is going. Crossing the threshold is the only moment that asks anything of you.
The root zone is 62 mm below full for this maize, out of 145 mm the soil can hold. The crop can comfortably use 96 mm of that before it starts working for water, and the forecast adds none.
How it works
The app holds you to a repeatable frame — tilt, distance, exposure — so this week's photograph is comparable with last week's rather than merely similar.
Vegetation indices and canopy cover are computed from the pixels. A converted consumer camera never yields a figure labelled NDVI; the app knows what your sensor can honestly produce.
Cover feeds the crop coefficient, the coefficient feeds the water balance, and the balance runs forward on the forecast to say when depletion crosses the threshold.
The result is written as a sentence with a date in it, citing the observations it came from. If the underlying numbers are thin, it says that instead of guessing.
Pricing
The water balance, the indices and the daily advisory cost us compute rather than a per-call fee, so they are not rationed. Only disease identification calls a specialist model on your behalf, and that is what a credit buys.
5 identifications, unlimited fields, and irrigation advice that does not stop when the credits do.
$29.99 · $1.20 each
25 identifications. A first season's worth of problem spots.
$79.99 · $0.80 each
100 identifications. What a weekly walk of the fields actually uses.
$249.99 · $0.62 each
400 identifications, across as many farms as you cover.
Credits do not expire while the account is open, and they belong to the organisation rather than to whoever bought them.
Questions
No, and it is not built to. It measures what a camera and a weather record can measure, and it is explicit about what it cannot see — soil water is not visible, which is exactly why the water balance is modelled rather than read off a leaf. An agronomist reviews the identifications the model is unsure about.
Every photograph submitted for identification is sent to Kindwise, a third-party classifier, because that is how the answer is produced. Imagery is stored in the European Union; the application and database run in Türkiye. The privacy notice sets out both, along with everything else we hold.
114 for measurements and disease identification — that is a general crop-health model rather than a per-crop table. Irrigation advice covers 68 of them, the crops FAO-56 publishes complete coefficients for, including maize, wheat, cotton, tomato, sunflower, olive, citrus, apple, pistachio and walnut. The app tells you which you are getting before you record a crop, rather than leaving you waiting for a figure that will never come.
A phone is enough. The app records tilt, exposure and focus so repeat captures are comparable, and it refuses to label a figure NDVI when the sensor cannot honestly produce one. Calibrated and multispectral sensors are supported where you have them.
Yes. Captures queue on the device and upload when you are back in range, which is the normal case rather than the exception. Results you have already seen stay readable offline.
Tell us, and it goes to review. A correction supersedes the machine answer rather than overwriting it, so the record still shows what was originally said and when it changed. Corrections are what improve the model.
Add a field, walk it once, and see what the balance says by the morning. It takes a photograph and a planting date — everything else is derived.
Contact
Questions about a crop we have not listed, a contractor arrangement, or how a figure was arrived at — write to us and a person answers.
[email protected]