I built FoodResQ, a JavaScript repository on my GitHub, to connect restaurants, bakeries, and cafeterias across Tamil Nadu with NGOs and volunteers so surplus food can be redistributed before it expires. The README frames the project around 1.3 billion tons of food wasted annually, 828 million people facing hunger worldwide, 30% of food produced being wasted, and Tamil Nadu festival and hotel surplus going unclaimed daily. The repository started and was last updated in 2026-10. My GitHub profile lists the project: my github profile. The live website is https://food-res-q-azure.vercel.app/ and the repository is https://github.com/ghkrknkp/FoodResQ.
- The README describes seven AI engines: Vision AI identifies Tamil Nadu dishes from photos, estimates weight and portions, checks freshness and allergens, and fills listing forms; the Tamil and English voice assistant uses Web Speech API (
ta-IN/en-IN) to extract food name, quantity, category, expiry, and location, with per-field confidence scores; the climate-aware expiry engine uses district profiles for Namakkal, Chennai, Madurai, Ooty, and Coimbatore to adjust safe windows for temperature and humidity, advise on storage, and trigger emergency SOS when the window is < 2 hours; the NGO matchmaker pairs donations by capacity with orphanages, shelters, and old-age homes, plans multi-stop routes with distance and ETA, and describes 15+ NGOs across Namakkal, Salem, and Dharmapuri; the FSSAI safety guard evaluates the 5°C–60°C bacterial danger zone, scores hygiene compliance, and provides a QR Safety Pass with Safe, Warning, and Danger indicators; the ESG engine calculates avoided CO₂ at 0.52 kg CO₂ per kg rescued, estimates water saved, and produces a downloadable Green Sustainability Certificate; and the kitchen-waste advisor uses weekday demand patterns, rainy-day footfall predictions, and the Tamil festival calendar to advise pre-cooking quantities.
The README separately describes Demand & Peak Timing Insights using 30-day MongoDB claim rates, a peak-claim-hour heatmap, and category success-rate analysis. The API reference also lists a listing assistant that enhances titles and tags, as well as endpoints for the seven named engines and demand insights. The README calls the suite “7” engines while describing demand insights separately and listing nine AI endpoints; the project structure says there are 8 AI panel components, and it does not explain how these counts relate.
I document the stack as React 18, Vite 5, Tailwind CSS 3, React Router v6, and Recharts on the frontend; Node.js, Express.js, Socket.IO, JWT, bcrypt, and Multer on the backend; and MongoDB with Mongoose. Socket.IO provides real-time communication. The README describes the AI engines as simulated, with Gemini API integration ready, and places live Gemini integration in Phase 3. Phase 1 (core platform) and Phase 2 (AI engines, analytics dashboards, and Tamil Nadu explorer) are marked complete; Phase 3 lists React Native, WhatsApp notifications, live GPS, and live Gemini API work.
The documented deployment separates the React + Vite frontend in client/ on Vercel from the Express API in server/ on Render, with MongoDB Atlas. Render uses root directory server, npm install, npm start, Singapore, and PORT=10000; Vercel uses root directory client, npm run build, and output directory dist. The frontend API URL is configured as VITE_API_URL=https://foodresq-api.onrender.com/api; the deployment instructions then set Render CORS to the live frontend origin, https://food-res-q-azure.vercel.app, and redeploy. The project structure identifies client/src/components/ai/ as containing 8 AI panel components, client/src/pages/ai/AiInnovationHub.jsx, client/vercel.json for SPA routing fallback, server/src/controllers/aiController.js for AI engines, server/src/routes/aiRoutes.js, render.yaml, and .env.example.
The API reference documents registration, login, and current-user profile routes; public food listing and detail routes; and restaurant-only food creation, update, and deletion for the owning restaurant. Its AI routes cover vision analysis, voice parsing, expiry calculation, NGO matching, safety verification, ESG reporting, waste forecasting, listing assistance, and demand insights. The repository structure also identifies client/vite.config.js.
git clone https://github.com/ghkrknkp/FoodResQ.git
cd FoodResQ
npm install && cd client && npm install && cd ../server && npm install && cd ..
cp .env.example server/.env
# Edit server/.env — set MONGODB_URI and JWT_SECRET
npm run dev
# Open http://localhost:5173