Farm Produce Marketplace Web App Connecting Farmers and Local Buyers

A farm produce marketplace web app connecting farmers and local buyers removes the middleman from the first mile of farm trade. Farmers list crop, variety, quantity in quintals and price per quintal; buyers search by crop or district, see stock next to a mandi reference price, negotiate in chat, and order against a delivery slot. Small farmers sell through commission agents or mandis, where transport, waiting time and commission eat thin margins, while small buyers — kirana shops, tiffin services, hostels — overpay for the same produce through wholesalers. Neither side can easily find the other. This project builds that meeting point: a searchable marketplace with dual roles, a negotiation simulation modelling real haggling, and a farmer dashboard for listings and orders. The demo build ships 16 sample listings across 6 crops with a working order flow, demonstrating the full buyer journey end to end.

A farmer creates a listing — crop, variety, quantity in quintals, asking price, harvest date, village — and it appears on the marketplace home as a photo card with a mandi reference price. A buyer searches for the crop, filters by district, and opens the listing to see the full specification sheet. Price is where the two sides meet: the buyer opens the negotiation chat, types an offer per quintal, and the simulated farmer accepts near the ask, counters within a margin, or declines below a floor tied to the mandi rate. Once the price is agreed, the buyer selects a quantity and one of six delivery slots across three days and confirms; the order appears in the farmer dashboard as pending and the listing's available stock drops. The farmer accepts or declines the order, completing the loop from harvest listing to confirmed sale.

1. A farmer posts a listing: crop, variety, quantity in quintals, price per quintal, harvest date, village/district and a produce photo (1 photo per listing in the demo build).
2. The listing renders as a marketplace card with the mandi reference price beside the asking price, anchoring the buyer's expectation before negotiation starts.
3. A buyer searches or filters by crop and district, opens the detail page, and negotiates in chat — the simulated farmer accepts offers at 95% of the ask or above, counters in the 80–95% band, and declines below the mandi floor.
4. The buyer confirms quantity and picks one of 6 delivery slots (3 days × 8 AM–12 PM / 2 PM–6 PM); available stock reduces immediately and the order is recorded.
5. The order lands in the farmer dashboard's incoming-orders panel, where the farmer accepts or declines it; accepted orders count toward the committed-quintals statistic.
6. Online payment stays optional: the base build confirms orders without payment; UPI/card collection via Razorpay can be added as a module.

Tech stack:

• React.js (marketplace + farmer dashboard UI)
• Node.js + Express (REST API)
• MongoDB (listings, orders, users, chat)
• JWT (buyer/farmer authentication)
• Socket.IO (real-time negotiation chat)
• HTML5 · CSS3 · JavaScript (working single-file demo)
• Razorpay (optional online payments)
• Git (version control)

FarmSetu ships four working capabilities that the student can demonstrate live. The marketplace home lists every farmer listing as a card with a real produce photo, quantity in quintals, a price that toggles between per-quintal and per-kilo, and a nearby mandi reference price that anchors the negotiation before it starts. The listing detail page pairs a specification sheet with a negotiation chat: the buyer types an offer and the simulated farmer accepts, counters or declines using mandi-aware rules. The order flow converts agreement into action — the buyer picks a quantity, chooses one of six delivery slots across three days, and the order lands in the farmer's dashboard. Finally, the farmer dashboard gives the seller a profile, statistics and two working panels: a listings table with price editing, and an incoming-orders panel with accept and decline actions.

Implemented in the demo build:

• [Farmer listings with produce photos] Farmers create listings with crop, variety, quantity, harvest date, village and district; in the demo build each listing carries one real produce photo (16 listings across 6 crops).
• [Buyer search, crop filters and sorting] Buyers search by crop, farmer, village or district, filter by crop chip, and sort by price or quantity — all client-side and instant in the demo.
• [Quintal-to-kg price toggle] Every price toggles between ₹/quintal and ₹/kg (1 quintal = 100 kg), so retail buyers who think in kilos and farmers who quote in quintals negotiate without unit confusion.
• [Mandi-anchored negotiation chat] Each listing shows a nearby mandi reference price, and the simulated farmer accepts, counters or declines offers relative to that anchor — modelling real haggling instead of free-form chat.
• [Order placement with delivery slots] Orders capture quantity and one of 6 delivery slots (3 days × morning 8 AM–12 PM / evening 2 PM–6 PM); confirming reduces available stock and routes the order to the farmer's dashboard.
• [Farmer dashboard] Profile with buyer ratings, four statistics tiles, a listings table with working price editing, and an incoming-orders panel with accept and decline actions that change order state.

Optional / future scope (not in the base build):

• [Online payments] Payment collection is intentionally excluded from the base build; UPI and card collection through the Razorpay payment gateway is an optional add-on, with orders marked paid before farmer acceptance.

• Complete source code: React frontend, Node.js + Express backend, MongoDB schemas
• Working single-file demo with 16 sample listings and the full order flow
• Database setup guide and REST API documentation
• Negotiation-chat simulation rules, documented and editable
• Project report PDF (background, methodology, architecture, screenshots)
• PPT presentation for final review
• Viva Q&A preparation document (marketplace logic, auth, real-time chat, trade-offs)