RAG Play - Retrieval-Augmented Generation System

A modern RAG (Retrieval-Augmented Generation) system built with React frontend and Python backend, deployed on Vercel.

Features

• Document Upload: Upload PDF, TXT, and DOCX files with metadata
• Semantic Search: Advanced vector search using Pinecone and BAAI embeddings
• AI-Powered Answers: Generate contextual answers using OpenAI or Groq
• Modern UI: Clean React interface with real-time search
• Serverless: Fully deployed on Vercel with Python serverless functions

Tech Stack

Frontend

• React 19 with Vite
• Modern CSS with responsive design
• PDF.js for document processing

Backend

• Python serverless functions (Vercel)
• Pinecone vector database
• Hugging Face Inference API for embeddings
• OpenAI/Groq for text generation

Quick Start

1. Clone the repository

   git clone <repository-url>
   cd RAG-Play

2. Set up environment variables Create a .env file in the root directory:

   PINECONE_API_KEY=your_pinecone_key
   HUGGINGFACE_API_KEY=your_hf_key
   OPENAI_API_KEY=your_openai_key
   

3. Install dependencies

   npm install

4. Run development server

   npm run dev

5. Deploy to Vercel

   vercel --prod

API Endpoints

• /api/search - Document search and answer generation
• /api/upload - Document upload and processing
• /api/docs - List all uploaded documents

Environment Variables

See api/README.md for detailed environment variable configuration.

License

MIT License