eco AgriSync
Architecture AI Engine About Stack Dashboard Contact

The field finally speaks data.

AgriSync · Version 0 — the first scope we build

AgriSync watches a single cabbage field in Uttar Pradesh. One cheap box in the ground, plus free satellite and weather data, spots a disease 12 to 72 hours before it shows on the leaves — so the farmer can spray in time, not too late.

Dashboard arrow_forward
Explore
Listening · Live

The field has a pulse.

Leaf wetness, air, sky and soil fold into one living signal. AgriSync watches it breathe — and feels the field slipping toward disease before a single spot shows on the leaf.

sensors Field box satellite_alt Satellite cloud Weather

One living signal · updated every day

Why Timing Is Everything

By the time you can see it, the spray window is half closed.

Disease control is a race against the clock. AgriSync moves the starting line 12 to 72 hours earlier — from the day spots appear, to the days before.

  1. notifications_active

    12–72 h early

    AgriSync raises the risk flag — before anything shows on the leaf. This is your spray window.

  2. visibility

    Symptoms appear

    The first spots become visible — when most tools begin.

  3. timer_off

    Too late

    The spray window is half closed and yield is already lost.

How AgriSync v0 Works

Five simple parts. One early warning.

sensors 01 / The Field Box

A Small Box in the Field

One cheap box sits in the field and measures what a disease needs to grow: how long the leaves stay wet, the air temperature, the humidity and the rain. It runs on a small solar panel, so it keeps working even when the power does not.

satellite_alt 02 / Satellites

Free Eyes in the Sky

Free Sentinel-2 photos check the crop's health every 2 to 3 days, and Sentinel-1 radar sees right through monsoon clouds. Together they give a wider view of the whole field, not just the spot where the box sits.

cloud 03 / Weather

Free Weather Data

Free weather history and a daily forecast for that exact field, from NASA POWER, ERA5-Land and Open-Meteo. The forecast is what lets the system warn ahead of time instead of after.

developer_board 04 / The Computer

One Cheap Computer at Home

A small computer at home pulls all the data together and runs the models. No cloud, no office and no big bills — it does its work quietly each day on its own.

database 05 / The Record

The First Local Disease Record

Every reading is saved in one small file on the computer. Over a season this grows into the first real record of cabbage disease for this crop and this corner of India — something no one has ever collected here.

monitoring The Output · The Warning

A Simple Dashboard

The disease-risk warning and the season's record show up on one simple dashboard. There is no farmer app, Hindi messaging or payments in v0 — those are planned for a later version, not part of this first build.

How It Thinks · v0

Learn healthy.
Catch the change.

Think of a smartwatch. It learns what your normal looks like and warns you when something drifts, before you feel sick. AgriSync does the same for a cabbage field. It learns what a healthy field looks like and quietly notices when the field starts slipping toward disease — usually before any spot shows on a leaf.

12–72h

How early it warns

2

Diseases watched in v0

70–75%

Target accuracy, with error bars

1

Field where it starts

satellite_altSatellite
cloudWeather
show_chartDisease risk
Live
12–72h early warning
Step 1 · Weather Rules

The Dependable Floor

Plant scientists already know the weather each disease needs. AgriSync runs those proven rules against the live readings and the forecast. This works on day one with no training data — about 60 to 70% right on its own.

Step 2 · Learn Healthy

Spot the Drift

The system learns the normal pattern of a healthy field from the box and the satellites, then flags when the live field drifts away from normal. The clever part: it never needs photos of sick plants to do this.

The Check · Is It Really Disease?

Drift Plus Weather

A drift on its own only means "something is off" — it could just be dry or hungry soil, not disease. So the system checks the weather too. A drift that matches the disease weather is a confident warning.

The Honest Bit · How Sure

A Risk, Not a Promise

The warning is given as a risk that rises over a day or two, with a clear margin of error — never a fake-exact "disease at 3pm". Like a smartwatch, it will sometimes raise a false alarm, and it is honest about that.

Crop Intelligence

Built for Banda Gobi

Cabbage disease, tuned to one field on the Indo-Gangetic Plain

AgriSync v0 starts with the two diseases that cost a Harpur cabbage farmer the most. Two other common ones are mapped out but parked for a later version, so v0 stays small enough to actually finish.

v0 · Watched Now

Alternaria Leaf Spot

Brown spots with rings, like little targets. It strikes when days are warm (20–30°C), the air is very damp, and leaves stay wet for 9 hours or more. AgriSync warns 12 to 72 hours ahead.

v0 · Watched Now

Downy Mildew

Yellow patches on top of the leaf, with grey fuzz underneath. It loves cool nights (8–16°C), damp air and morning fog. The worst risk is in October and November.

Later version

Black Rot

Yellow V-shapes at the leaf edge that turn black. It spreads in warm, wet weather. Already mapped out, but saved for a later version to keep v0 small.

Later version

Soft Rot

Mushy, smelly rot that comes with waterlogged soil and in storage. The fix is better drainage and harvest timing. Saved for a later version.

The Gap

Everyone else looks back.
AgriSync looks ahead.

Plenty of tools touch this problem. None combine pre-symptom prediction, one crop, one microclimate, and cheap shared hardware for a smallholder. That farmer sits below the economic radar of the big players — and that gap is the opening.

Per-plot hardware

Fasal, Fyllo

Accurate sensors on every plot

Costs tens of thousands per plot; tuned for vineyards and orchards, not a quarter-acre cabbage field

Satellite platforms

Cropin, Satsure, Farmonaut

Cheap satellite data sold to large buyers

Generic across crops; the farmer is a data point, not a user

Hyperspectral imagery

Pixxel

High-detail imagery from space

Sells pixels, not decisions

Photo diagnosis

Plantix

Identifies disease from a leaf photo

Reactive — only works once the symptom is already visible

Government extension

Public advisories

Trusted, free advisories

Slow, district-level and one-size-fits-all

eco

AgriSync v0

Predict, don’t react

checkPredicts before symptoms appear
checkOne crop, one microclimate
checkOne shared low-cost box per cluster
checkFarmer-first, not corporate-first

Honest caveat: pre-symptom prediction already exists in hyperspectral lab and greenhouse research. What does not exist is a deployed product doing it for smallholders with one crop, one microclimate and cheap hardware. The honest claim is “no deployed smallholder product does this,” not “nobody has ever tried it.”

The Resolve

Built from one village field,
built honestly.

cottage

Tuned to One Real Field

Built for one real cabbage field in Harpur, eastern Uttar Pradesh. Every weather rule and disease window is tuned for that farmer's field and his neighbours, not a generic crop.

language

The Real Prize Is the Record

The biggest result is not the prediction — it is the record. AgriSync is building the first-ever local record of cabbage disease for this crop and region. That record simply does not exist anywhere today.

handshake

Honest, Not Hyped

The satellite images, weather and tools are all free or open. v0 keeps its promises small and clear: it will sometimes raise a false alarm, and every weak point has a backup plan written down.

sensors

In the field

Solar box · leaf wetness, temperature, humidity, rain

developer_board

At home

One Raspberry Pi 5 pulls the data and runs the models

monitoring

The output

One simple dashboard · risk curve and season record

Village Field #01

Field intelligence,

redefined.

The whole thing runs on one small computer.

No cloud bills and no big setup — just free satellite data overhead, one small box in the field, and one device on the wall.

A solo project from a village in India — built first by one person, so a research university can make it rigorous and bring it to many farms.

Under the Hood

The AgriSync Tech Stack

developer_board Raspberry Pi 5
terminal Python 3.11
schedule systemd
database SQLite
storage DuckDB
api Flask
model_training XGBoost
public ERA5-Land
satellite_alt Sentinel-1 / 2
wb_sunny NASA POWER
cloud Open-Meteo
water_drop Leaf Wetness Sensor
thermostat BME280 Sensor

Want to follow the build?

AgriSync v0 is being built in the open for one cabbage field this season. Get in touch if you grow cabbage, work with satellites or machine learning, or want to help prove it properly.

CONNECT Connect on LinkedIn
API Reference Hardware Specs Deployment Guide