SmartServe is an AI-powered food waste optimization system that predicts surplus, recommends low-waste menu alternatives, and computes optimal portion sizes.
It is built using React, FastAPI, Supabase, and ML tools including Prophet, SentenceTransformer, PuLP, and the Gemini API for intelligent suggestions.
Live Demo: https://smartserveproject.netlify.app/
Institutional kitchens often overprepare food due to inaccurate forecasting and intuition-based planning.
SmartServe provides a data-driven solution by combining forecasting, menu analysis, optimization, and LLM-powered insights into a single platform with an admin dashboard, prediction service, and basic ordering app.
- Browse menu with nutritional information
- Place real-time food orders
- Track order status
- Responsive UI built with React and Flutter
- Live order dashboard
- Daily statistics: completed, pending, total, predicted demand
- Wastage analytics and item performance
- Menu management with availability control
- Automatic demand predictions using historical data
- FastAPI ML service for:
- Demand forecasting (Prophet / Regression models)
- Menu similarity analysis (SentenceTransformer)
- Portion optimization (PuLP)
- Gemini API for paragraph-style insights and actionable suggestions
Institutional canteens face challenges such as:
- Overproduction leading to food wastage
- Underproduction causing shortages
- No forecasting or data-driven planning
- No system to track peak hours or popular items
SmartServe solves this by providing:
- Real-time demand visibility
- Forecast-based planning before cooking begins
- Reduced wastage and better resource allocation
- A smooth digital ordering experience
- React.js
- Tailwind CSS
- ShadCN UI
- Flutter (basic ordering app)
- Zustand for state management
- FastAPI (ML backend)
- Gemini API for suggestion generation
- Prophet, SentenceTransformer, PuLP
- Supabase (PostgreSQL + Auth)
- Real-time listeners for orders
- Netlify — frontend
- Render — FastAPI backend
- Supabase — hosted database
- Users place orders via web or Flutter app
- Orders stored in Supabase
- Admin receives real-time updates
- FastAPI consumes:
- Historical orders
- Item popularity
- Day type (weekday, peak, event days)
- Returns:
- Predicted demand
- Estimated wastage
- Suggested portion adjustments
- View current-day analytics
- Track completed vs pending orders
- Monitor expected demand
- Access AI-powered suggestions
- Frontend: Netlify
- Backend APIs: Render
- Database: Supabase (hosted)