🤖KS RAG Chatbot - PDF Question Answering System

A full-stack Retrieval-Augmented Generation (RAG) chatbot for uploading PDFs and asking questions about their content. Built with FastAPI, Next.js, MongoDB, Cohere embeddings, and Groq's Llama 3.3 70B.

✨ Features

• 📄 Upload up to 4 PDFs simultaneously with automatic chunking & vector embeddings
• 💬 Context-aware Q&A using semantic search and RAG
• 👤 JWT authentication with secure password hashing
• 🗂️ Multiple chat conversations with auto-generated titles
• 🎨 Modern dark UI with responsive design

🚀 Quick Start

Prerequisites

• Python 3.11+, Node.js 18+, MongoDB
• API Keys: Cohere & Groq

Installation

1. Clone & Setup Backend

git clone https://github.com/Krishna41357/KS-RAG.git
cd KS-RAG/server

python -m venv ragbot
source ragbot/bin/activate  # Windows: ragbot\Scripts\activate
pip install -r requirements.txt

# Create .env file
cp .env.example .env

Edit server/.env:

MONGO_URI=mongodb://localhost:27017/rag_database
COHERE_API_KEY=your_cohere_api_key
GROQ_API_KEY=your_groq_api_key
SECRET_KEY=your_secret_key_32_chars_minimum

Start backend:

uvicorn main:app --reload --port 9000

2. Setup Frontend

cd ../client
npm install

# Create .env.local
cp .env.example .env.local

Edit client/.env.local:

NEXT_PUBLIC_API_BASE_URL=http://localhost:9000

Start frontend:

npm run dev

Visit http://localhost:3000

📖 Usage

1. Sign up/Login at the homepage
2. Upload PDFs using the paperclip icon (up to 4 files)
3. Ask questions and get AI-powered answers with source citations
4. Manage chats from the sidebar

🛠️ Tech Stack

Backend: FastAPI, MongoDB, PyPDF2, Cohere, Groq, JWT
Frontend: Next.js 16, TypeScript, TailwindCSS, Lucide React

🔌 Key API Endpoints

Method	Endpoint	Description
POST	/users/register	Register user
POST	/users/login	Login user
POST	/upload	Upload PDFs
POST	/chats/{id}/messages	Send message & get response
GET	/chats	List user chats

📁 Project Structure

KS-RAG/
├── server/              # FastAPI backend
│   ├── main.py         # Entry point
│   ├── vectorstore.py  # PDF processing
│   ├── routes/         # API endpoints
│   └── .env.example
├── client/             # Next.js frontend
│   ├── app/           # Pages
│   ├── components/    # React components
│   └── .env.example
└── README.md

👨‍💻 Author

Krishna Srivastava
GitHub: @Krishna41357
Email: krishnasrivastava41357@gmail.com

📄 License

MIT License

---

Built with ❤️ by Krishna Srivastava