NetBhav — Mandi Opportunity Agent
Link to open source: https://github.com/ace-ify/netbhav
Link to Live Project: https://netbhav.vercel.app/
NetBhav is an AI-assisted decision-support tool that tells a farmer which nearby mandi actually puts the most money in their pocket — not just the one with the highest sticker price.
The problem (PS-02): Farmers pick where to sell using scattered mandi price boards. But a higher price often earns less once transport, commission and mandi fees are subtracted. That gap silently costs farmers thousands per harvest.
What it does: Take a farmer's crop + quantity + location → pull live mandi prices from multiple sources → compute transport + every deduction (commission, mandi fee, cess, hamali, spoilage) → rank nearby mandis by real take-home, and give one clear recommendation plus the ₹ gained over the naive choice. Example: for 50 quintal mustard near Lucknow, chasing the highest sticker price loses ₹9,556 — NetBhav catches it.
Why we built it: The honest math is simple but no farmer-facing tool does it transparently, in Hindi, on any phone. A pure-TypeScript, unit-tested engine does every rupee; the AI only handles language.
How it's useful — one profile, many doors:
- Web/PWA (Hindi-first) with a live map + real road routes
- Hindi voice — in-browser and via a real phone call (LiveKit + Twilio)
- WhatsApp/SMS — just message your crop and place
- Proactive alerts — it watches prices and calls/messages you when it's time to sell
- Admin/operator console for FPOs & govt with "farmer income surfaced" metrics
Highlights: multiple independent data sources (live Agmarknet mirror + data.gov.in + CEDA + WPI) with honest freshness badges, any-crop fuzzy search (Hindi/English/typos), FPO bulk-pooling mode, class-aware spoilage math, and full graceful degradation (runs with zero paid keys).
Links:
- Live demo:
FOR FARMER: https://netbhav.vercel.app/app
FOR ADMIN: https://netbhav.vercel.app/admin
- Github: https://github.com/ace-ify/netbhav
This build was uploaded as a hackathon project
