Built for Indian farms

Know what to plant, when to water, and what's wrong with your crop.

AGRO MIRAI reads a field's soil, weather, and satellite signals to ground every recommendation in real data, not a guess — and speaks it back in the farmer's own language.

Speaks
Crop advice ಬೆಳೆ ಸಲಹೆ పంట సలహా फसल सलाह
Scroll
Bellary district, Karnataka ICAR & state agriculture data ಕನ್ನಡ · తెలుగు · हिन्दी · English
What it does

Four advisories, one field visit

Every recommendation is generated from the farmer's own field data — soil samples, local weather, and satellite NDVI — not a generic lookup table.

Crop recommendation

Ranks all 22 supported crop types against FAO EcoCrop's own temperature, soil-pH, and soil-texture ranges — plus the field's real 12-month rainfall and a sowing-month calendar. A regional-suitability check flags any pick that's agronomically unusual for the farmer's district, instead of silently trusting the model.

Irrigation advisory

A daily FAO-56 root-zone water balance — Hargreaves-Samani reference evapotranspiration (chosen over Penman-Monteith, since a real farm doesn't reliably report solar radiation or wind), crop coefficient by growth stage, root depth, and soil water-holding capacity. The same numbers set both the recommended depth in mm and the urgency — never two models that can disagree.

Disease risk detection

A MobileNetV2 CNN trained on the PlantVillage dataset reads an uploaded leaf photo, when one's available. When it isn't, a weighted rule-based score over humidity, rainfall, temperature, and NDVI trend fills in — and if the CNN call itself fails, that same rule-based score catches it. Never a dead end.

Voice, in four languages

Every advisory translates and speaks back in English, Kannada, Telugu, and Hindi (IndicTrans2 + AI4Bharat ASR/TTS, Piper for English and Hindi) — and login is just name, phone, and an OTP, so there's no password for a farmer to forget.

How it works

From field data to a farmer's advisory

The same pipeline runs behind every advisory type — only the model at the processing step changes.

Input

Field & environment data

  • Soil N/P/K, pH, moisture
  • Daily weather (temp, humidity, rainfall)
  • Satellite NDVI (Google Earth Engine, with a local NDVI cache fallback)
  • Optional leaf photo upload
Processing

Feature build & model inference

  • Feature vector assembled per field, per request
  • EcoCrop suitability ranking · FAO-56 water balance · disease CNN or rule score
  • Plain-language rationale attached to every result — the numbers the rule used, not a black box
Output

Advisory delivered

  • Plain-language summary, severity-tagged
  • Mobile app (React Native / Expo) & admin dashboard
  • Spoken aloud on request, in the farmer's language

Every figure on this page comes from a real, committed evaluation report — nothing hand-faked.

In the farmer's pocket

One app. Every field, every advisory.

A React Native / Expo mobile app puts all four advisories, voice playback, and photo-based disease scans on one screen a farmer already knows how to use — no password, just name, phone, and an OTP.

  • Name + phone + OTP login — nothing to remember
  • Live field readings, translated and read aloud
  • Camera-based leaf scan for disease risk, on the spot
Download the Android app (APK, 39.5MB)

Direct install — no Play Store yet. Android will ask to allow "install from unknown sources". Release notes & checksum.

AGRO MIRAI Live
North Plot · English ಉತ್ತರ ಪ್ಲಾಟ್ · ಕನ್ನಡ నార్త్ ప్లాట్ · తెలుగు नॉर्थ प्लॉट · हिन्दी
Recommended crop
Cotton
82% confidence
Irrigation depth
12 mm
Moderate urgency
Disease risk
Moderate
Scout within 2 days
Measured, not claimed

Real numbers, not marketing claims

Where a figure comes from a trained model's held-out evaluation, we say so. Where it comes from a rule or a real dataset instead, we say that too — crop and irrigation run on transparent methods, not a black-box accuracy score.

0%
Disease CNN validation accuracy
MobileNetV2 on PlantVillage, 38 classes — the only trained model in the pipeline
So what: this is the only number on this page that comes from a trained model's own held-out score — every other figure below is a real count, not a model claiming confidence in itself.
0
Crop types ranked
FAO EcoCrop suitability data, not a trained-model percentage
So what: ranked against real agronomic tolerance ranges, not a lookup table.
0
Crops verified regionally correct for Bellary
Cross-checked against ICAR-CRIDA & state agriculture data, not the model's own opinion
So what: a second, independent check the model can't override.
ಕನ್ನಡ తెలుగు हिन्दी English
Indian languages supported
Translation, speech-to-text & text-to-speech, all four
So what: a farmer never has to read English to use this.