Anna Bazaar
Anna ka rang, Zindagi ka sang.
The problem Anna Bazaar solves
The Problem Anna Bazaar Solves
Small and marginal farmers in India face a rigged system. They sell through layers of middlemen who quietly eat into their margins, often don't see real-time mandi prices, and end up doing distress sales just to clear loans.
Existing agri platforms help a bit, but many are region-locked, crash on low-end phones, or still route trades through agents instead of giving farmers direct buyer access. Add language barriers, patchy digital literacy, and climate shocks, and you get a system where the people growing our food have the least bargaining power.
What Anna Bazaar Solves
Direct farmer-to-buyer connection eliminating middlemen through transparent listing and chat-based negotiation.
Real-time price protection via our AI-driven Dynamic Mandi Price Engine that suggests safe price bands and labels offers as LOW, FAIR, or GENEROUS.
MSP enforcement as a safety net — the engine never nudges farmers below Minimum Support Price, even if mandi rates drop.
Multilingual support in Hindi, Bengali, Marathi, Telugu, Tamil for farmers with low digital confidence.
AI-powered crop listing using computer vision to auto-fill crop type and quality grade, saving time.
Secure digital payments and logistics integration tailored for agriculture commerce.
Cartel detection alerts when the system notices repeated suspiciously low bids from the same buyer cluster.
How It Makes Things Easier and Safer
Saves Time: AI automation reduces crop listing time by over 70%.
Increases Income: Removes intermediaries and protects farmers from lowball offers.
Builds Trust: Real-time mandi data from credible government sources (Agmarknet, official MSP lists) with clear timestamps.
Empowers Decisions: Live price bands, weather data, and fairness scores help farmers negotiate confidently.
Enhances Security: KYC verification, encrypted data, and secure payment flows protect transactions.
Use Cases
Small farmers can sell directly to buyers at transparent, fair rates backed by real data.
Rural producer groups can aggregate produce and negotiate bulk deals efficiently.
Retailers and urban consumers gain access to fresh, traceable farm produce with full transparency.
Agri-businesses and cooperatives can leverage platform data for better market insights.
Challenges we ran into
Building Anna Bazaar in 24 hours meant hitting walls repeatedly. Here's what nearly broke us:
The Midnight Data Disaster
At 12:30 AM, we wired up the Agmarknet API and saw ₹0.00 for potatoes across 47 mandis. Panic mode. After an hour of debugging, we realized: those mandis were just closed on December 27th. We added fallback logic to use yesterday's closing price with a stale-data warning.
Lesson learned: Real-world data is messy, and your algorithm needs to handle it gracefully.
Formula Wars
We spent 40 minutes arguing whether the demand factor should adjust prices by 5%, 10%, or 15%. Finally, someone said: "If we can't explain this to a farmer in 10 seconds, it's too complex." We settled on 10% adjustments with clear labels (WEAK DEMAND / NORMAL / HIGH DEMAND).
Goal: Explainable fairness, not a finance PhD thesis.
The 2 AM Type Mismatch From Hell
First time we connected
mandiPriceService.ts
to React components, the UI went blank. No errors. Just nothing. Turned out: the service returnedpriceData
as a number, but the component expected an object with avalue
field. Took 30 minutes to find because our eyes were half-closed.Debugging at 2 AM is purgatory.
Ensuring Mobile-First Performance
Making AI-powered quality checks work reliably on low-end Android phones with intermittent 2G connectivity required using TensorFlow Lite for on-device predictions and aggressive caching strategies.
Building Multilingual Trust
Integrated Google Translate for 22 Indian languages, but had to carefully test microcopy like "This offer is ₹6/kg below your safe price range" to ensure it stayed clear and empathetic across translations.
Energy Dips and Team Dynamics
Around 4 AM, git conflicts piled up (three people editing
App.tsx
simultaneously). We snapped at each other about variable naming. A chai break and one successful test run restored the mood.Hackathons test team dynamics as much as technical skills.
Last-Minute Microcopy Polish
With 2 hours left, we rewrote every error message from
PRICE_BELOW_THRESHOLD
to human-readable text like "This offer is ₹6/kg below your safe price range."That small change made the product feel empathetic instead of robotic.
Technologies used