A full-stack web application that helps students upload PDF documents and receive AI-generated summaries and multiple-choice questions using HuggingFace pre-trained language models. (Project is still under development)
- React.js (Vite)
- Tailwind CSS
- Firebase Authentication (Email/Password)
- Axios for API calls
- Node.js
- Express.js
- Multer (file uploads)
- pdf-parse (PDF text extraction)
- Axios (HuggingFace API calls)
- Winston (logging)
- Helmet (security)
- Compression (performance)
- Express Rate Limit (DDoS protection)
- Express Validator (input validation)
- MongoDB (Mongoose ODM)
- Connection pooling
- Optimized indexes
- HuggingFace Inference API
- Circuit breaker pattern
- Retry logic with exponential backoff
- Models:
facebook/bart-large-cnn(summarization)google/pegasus-xsum(summarization)- Multiple fallback models
- User authentication with Firebase
- PDF upload with drag & drop
- AI-powered text summarization
- Automatic MCQ generation (5 questions per document)
- Session history
- Dark and light themed modern UI
- Responsive design
- Protected routes
- Error handling