Behind a simple screen — "when to spray, when to enter the field" — runs an engine built on the same science institutions pay for: 130+ peer-reviewed agronomic models, dozens of satellite streams fused into one view, decisions computed for your specific parcel, hourly, across the whole season. It is not a skin over a weather forecast. It is a system.
Most "farming apps" show you weather and NDVI. Niezbędnik makes decisions — 60+ distinct ones, each for a specific parcel and its growth stage:
One crop is not one model. Wheat, oilseed rape, potato, pea, apple, onion — each has its own phenology, its own diseases, its own pests and its own thresholds. The system does not lay one lookup table over everything: for each crop group it computes the risks that genuinely apply to it, at the BBCH stage matched to its variety and sowing date.
Septoria, fusarium head blight (DON risk), mildew, rusts, tan spot — infection windows tied to heading stage and leaf-wetness hours. Nitrogen and growth-regulator thresholds per stage.
Phoma stem canker, sclerotinia stem rot, dark leaf spot; pollen beetle and weevil flights computed on temperature sums. The flowering window drives protection timing.
Late blight (BLITECAST / Hutton model), early blight, Colorado beetle; cercospora leaf spot and ramularia in beet. Irrigation thresholds during tuber and root set.
Ascochyta, downy mildew, aphids; plus a nitrogen credit for the next crop in the rotation. High drought sensitivity during flowering.
Apple scab (Mills model, wetness hours), fire blight, codling moth on temperature sums. Blossom frost risk computed per phenological stage.
Onion downy mildew, carrot alternaria, root flies; precise irrigation thresholds and narrow spray windows with drift control.
The same weather event means something different for each crop — and the system tells them apart. Rain at wheat heading is a fusarium risk; the same rain at rape flowering is a sclerotinia risk; in the orchard it is a scab window. Computed for your parcel and its crop — not for a region, not for a municipality.
The difference between an "app" and a system is measurable here. Each recommendation rests on a peer-reviewed publication and is calibrated to the Polish agricultural climate. These are not thresholds copied from a leaflet — they are mechanistic models, the same families science uses:
Septoria, fusarium head blight, late blight, sclerotinia, phoma — each with its own infection window, latency period and wetness kinetics. Built on established phytopathology literature.
A physical development and biomass model corrected on the fly by satellite NDVI — a filter joins physics with observation, so the growth stage does not drift away from reality.
A full soil-water model with effective rainfall, crop- and stage-dependent evaporation, and capillary rise from groundwater — across eight soil-texture classes.
The model families this builds on (publicly citable method families, without disclosing our implementation):
A single sensor lies: cloud blocks the optics, radar sees no colour, microwaves see no leaf. So we do not look with one eye — we fuse dozens of streams and weight them by uncertainty, so each covers the weakness of the others:
Optical (condition, vegetation indices), SAR radar (lodging, mowing, moisture — through cloud), thermal (water stress before the eye can see it). Together — a picture no single satellite gives.
Root-zone moisture from microwaves, surface water from the balance, deep water from gravimetry. We combine it with a local model — from dew on the leaf to the reserve underground.
Every stream is fitted to your parcel — the weather point nearest its centre, terrain from a local DEM, the boundary from the land register. Not a "region", not a "municipality" — a specific field.
Commercial agricultural intelligence of this class typically costs tens of euros per hectare per year — and it tends to reach large farms and corporations. We give the core to the farmer free of charge, and that is not charity but architecture:
Real fields test our models better than any laboratory. Each season of a farmer's decisions — what worked, what did not — helps calibrate the same models before they reach paying clients: insurers, banks, commodity firms. The farmer's free use is part of our validation. That is what lets us give it away.
What "free" means for you: the tool costs you nothing, and aggregated, anonymized field outcomes feed back into model validation. We do not sell your individual records; what flows back is the pooled signal of whether a model's call matched what happened in the field.
130+ peer-reviewed agronomic models and 60+ per-parcel decisions, all curated rather than generated — recomputed per parcel across the full season, not a static dashboard.
Works with no signal — the treatment log and field notes sync once back online. Per-country localisation: weather, crop calendars, language. A release is in preparation for 8+ European countries (roadmap, not yet live everywhere).
We show the sources, the confidence and the reason for escalation (e.g. lodging from radar raises fusarium risk). An alert appears only when the cost of treatment is below the expected loss — an economic threshold, not panic.
What we are careful about. These are models, and models can be wrong. Satellite passes are missed under persistent cloud; sensor fusion narrows but does not erase uncertainty; phenology and disease thresholds are calibrated to a region and can drift for an unusual season or a new variety. We surface confidence alongside every call and flag when an input is stale or missing rather than guessing. The system supports a decision — it does not replace the agronomist on the ground.
Strip away the engineering and three things remain. Niezbędnik is not a weather app — it is a decision system, 130+ peer-reviewed agronomic models resolving into 60+ per-parcel calls, recomputed hourly across the season. It is free because real fields and full seasons of farmer decisions validate those same models before they reach paying institutional clients; the farmer's use is part of the engine, not a cost it carries. And it is one screen on a larger Earth-observation system — Niezbędnik Rolnika shares its core with products built for banks, insurers and local government. The same science, a different screen.
It is live today, and the release is being prepared across 8+ European countries — the same toolkit, localised per country. Open it on your own parcel and let it earn the screen.