🌾 AgroLink – AI-Powered B2B Marketplace for Farmers & Buyers
🧠 Inspiration AgroLink was born from a simple observation: farmers have the produce, but no direct digital channel to reach serious buyers. Most lack access to modern tools, real-time price insights, and localized technology. We wanted to create a full-stack platform powered by AI, voice, GPS, and a beautiful real-time experience that connects farmers directly to retailers, restaurants, and exporters—in their own language.
🛠️ How We Built It We built AgroLink as a real-time, full-stack web application with scalable and modular architecture:
🔧 Tech StacK Frontend: React + TypeScript + TailwindCSS + Framer Motion Backend & DB: Supabase (PostgreSQL, Auth, RLS policies, Edge Functions) Authentication: Email/password and social (Google, Facebook, LinkedIn, Twitter)
APIs & Features: 🌍 Google Maps API – GPS tagging & location-based search 🧠 Mock ML via Python (FastAPI) – AI crop predictions 🎙️ Web Speech API – multilingual voice assistant (Hindi & English) 🛒 Product catalog with 65+ SKUs: fruits, vegetables, grains, dairy
🌟 Major Features AI Crop Predictor – suggests ideal crops and yield estimation Voice Assistant – allows users to search and list produce using voice commands GPS Farm Tagging – auto-detects farm location for authenticity Secure Transactions – cart + checkout + payment simulation Marketplace Search Bar – future upgrade: Google Search API integration Notification Center – real-time order alerts Language Translator – English ↔ Hindi translation (WIP)
🔍 Challenges We Faced Voice permission issues due to browser and mic access Supabase table errors (public.profiles, crop_predictions, etc.) from missing schema setup Mapping 65+ product images correctly from online sources Making the site responsive, fast, and accessible on 3G connections Syncing Supabase RLS policies with user roles (farmer, buyer) securely Adding animated, real-time UI without slowing down performance
✅ What We Accomplished Created a production-grade B2B agri-marketplace with authentication, real-time features, and working voice/GPS/AI modules Full working cart + marketplace + dashboard flow Intuitive UI and engaging animations using Framer Motion Schema designed with modular security and auto-updating tmestamps
📚 What We Learned How to fully leverage Supabase RLS + Auth + Functions in a real product Integrating AI + GPS + voice into a single smooth UX Importance of clean schema setup to avoid cascading 404 and RLS errors Fast debugging via real-time Supabase logs and network monitoring Building scalable UI components for marketplace filtering and product browsing
🔮 What’s Next Integrate real AI models with weather and soil APIs Deploy full UPI & Razorpay for live payments Add Push Notifications for mobile devices Expand to more Indian languages (Marathi, Telugu, Tamil) Launch native Android app with offline sync for low-connectivity zones Implement blockchain-based traceability for premium exports
Log in or sign up for Devpost to join the conversation.