Agri Warehouse
AI-checked grain intake for agricultural warehouses: a vision model checks every truckload, every bag gets a QR code, and certificates are signed so nobody can fake them.
- Role
- AI full-stack developer at Toshi Consulting
- Year
- Jul 2026
- Stack
- Next.js, React, Python, PostgreSQL, OpenAI API, Vision model
- Links
- Pilot site
The problem
Grain quality and quantity are checked on paper at warehouse gates. Wrong commodities, short weights and forged certificates are hard to catch, and nobody outside can verify where a bag has been.
How it works
What I built
- 01AI photo verification: every truckload is photographed and a vision model cross-checks it against the declared commodity in real time.
- 02Bag-level QR tracking from intake through storage to dispatch, with a public chain of custody anyone can scan.
- 03Digitally signed (Ed25519), publicly verifiable certificates, plus automatic fraud alerts on weight differences and vehicle patterns, in English and Hindi.
Try it yourself
Run a truck through the gate. The vision check compares the photo with what was declared, the weights are compared, and the certificate gets a real digital signature that breaks if anyone edits it.
Truck declares
Gate camera photo
Vision model seesWheat · 97% sure
Against the declarationmatches ✓
Weight difference0.33% ✓ within 1%
Everything checks out. Issue the certificate to sign it.
Runs in your browser with sample data. It follows the real product's rules; it isn't connected to it.
The result
Running as a pilot in Haryana at warehouse.toshiconsulting.com, for warehouse officers, suppliers and citizens checking their grain.
- QR code on every bag, gate to dispatch
- 1
- QR code on every bag, gate to dispatch
- of truckloads checked by the vision model
- 100%
- of truckloads checked by the vision model
- languages, English and Hindi
- 2
- languages, English and Hindi