🧠 Legal QA System with Fine-Tuned LLMs and RAG

A scalable Legal Question Answering (QA) system that combines fine-tuned open-source Language Models (LLaMA, Mistral) with Retrieval-Augmented Generation (RAG) to deliver accurate, explainable answers from 100K+ U.S. legal documents.

🌐 Live Web Application: Access the system through a modern React frontend with FastAPI backend, connecting to the LurisQA model on Hugging Face.

---

🚀 Quick Start

Option 1: Deploy to Render (Recommended)

1. Fork this repository
2. Sign up at render.com
3. Create a new Web Service from your GitHub repo
4. Add your HUGGINGFACE_API_KEY as an environment variable
5. Deploy! 🎉

Option 2: Local Development

# Clone the repository
git clone <your-repo-url>
cd Legal-QA-System-with-Fine-Tuned-LLMs-and-RAG

# Set up environment variables
cp backend/env.example backend/.env
# Edit backend/.env with your Hugging Face API key

# Run with Docker Compose
docker-compose up --build

# Or run manually:
# Backend: cd backend && pip install -r requirements.txt && python main.py
# Frontend: cd frontend && npm install && npm start

---

🏗️ Architecture

Frontend (React)

• Framework: React 18 with Create React App
• Styling: Tailwind CSS for modern, responsive design
• Icons: Lucide React for beautiful icons
• HTTP Client: Axios for API communication

Backend (FastAPI)

• Framework: FastAPI for high-performance API
• Model Integration: Hugging Face Inference API
• CORS: Configured for cross-origin requests
• Error Handling: Comprehensive error management

Model

• LurisQA: 8.03B parameter LLaMA-based legal model
• Hosting: Hugging Face Inference API
• Access: Via API key authentication

---

🧰 Tech Stack

Module	Technology
Frontend	React 18, Tailwind CSS, Axios
Backend	FastAPI, Python 3.11
Model Hosting	Hugging Face Inference API
Language Models	JurisPrae (LLaMA-based, 8.03B params)
Fine-Tuning	LoRA (via PEFT, bitsandbytes)
Embeddings	SentenceTransformers
Vector Store	FAISS
Deployment	Docker, Render, Vercel, Railway

---

📦 Features

Web Application

• ⚡ Real-time Chat Interface: Modern, responsive UI for legal Q&A
• 🔍 Semantic Search: Powered by RAG technology
• 💬 Chat History: Persistent conversation tracking
• 🎨 Professional Design: Legal-themed UI with accessibility features
• 📱 Mobile Responsive: Works on all devices
• ⚠️ Legal Disclaimer: Built-in disclaimers for legal compliance

Core System

• ⚡ 4× Memory Efficient: LoRA tuning reduces GPU footprint
• 🔍 Semantic Search: Real-time retrieval of relevant legal paragraphs
• 💬 Answer Grounding: LLM outputs cite actual case law & statutory references
• 🛠️ End-to-End Pipeline: From document ingestion to user query generation

---

🌐 Deployment Options

1. Render (Full-Stack)

• ✅ Automatic HTTPS
• ✅ Global CDN
• ✅ Auto-deploy on Git push
• ✅ Free tier available

2. Vercel + Railway

• ✅ Vercel for frontend hosting
• ✅ Railway for backend API
• ✅ Excellent performance

3. Hugging Face Spaces

• ✅ Perfect for model showcase
• ✅ Integrated with Hugging Face ecosystem
• ✅ GPU support available

4. Local Development

• ✅ Docker Compose setup
• ✅ Hot reloading
• ✅ Easy debugging

See DEPLOYMENT.md for detailed instructions.

---

🧪 API Usage

Ask a Legal Question

curl -X POST http://localhost:8000/api/ask \
  -H "Content-Type: application/json" \
  -d '{"question": "What are Miranda rights?"}'

Health Check

curl http://localhost:8000/health

Model Information

curl http://localhost:8000/api/model-info

---

🔧 Configuration

Environment Variables

# Required
HUGGINGFACE_API_KEY=your_huggingface_api_key_here

# Optional
HOST=0.0.0.0
PORT=8000
LOG_LEVEL=INFO

CORS Settings

Update backend/main.py with your frontend domain:

allow_origins=["https://your-domain.com", "http://localhost:3000"]

---

📊 Example Usage

Web Interface

1. Open the web application
2. Type your legal question
3. Get instant, AI-powered answers
4. View chat history and citations

Programmatic Access

import requests

response = requests.post("http://localhost:8000/api/ask", 
                        json={"question": "What are Miranda rights?"})
print(response.json()["answer"])

---

🔒 Security & Compliance

• ✅ API Key Protection: Secure environment variable handling
• ✅ CORS Configuration: Proper cross-origin request handling
• ✅ Input Validation: Pydantic models for request validation
• ✅ Error Handling: Comprehensive error management
• ✅ Legal Disclaimer: Built-in disclaimers for legal compliance

---

🚀 Performance

• Response Time: < 5 seconds for most queries
• Concurrent Users: Supports multiple simultaneous users
• Scalability: Horizontal scaling via Docker containers
• Caching: Static asset optimization via nginx

---

🤝 Contributing

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Add tests if applicable
5. Submit a pull request

---

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

---

🙏 Acknowledgments

• Hugging Face for model hosting and inference API
• The open-source community for the amazing tools and libraries

---

Ready to deploy? Check out the Deployment Guide! 🚀