README.md

metadata

title: CyberLegalAIendpoint
sdk: docker
app_port: 8000

CyberLegal AI - LangGraph Agent

Advanced cyber-legal assistant powered by LangGraph + LightRAG + GPT-5-Nano for European regulations expertise.

🏗️ Architecture

┌─────────────────┐    ┌──────────────────┐    ┌─────────────────┐
│   Client API    │────│  LangGraph Agent │────│  LightRAG Server│
│   (Port 8000)  │    │   (Orchestration)│    │  (Port 9621)    │
└─────────────────┘    └──────────────────┘    └─────────────────┘
                               │
                       ┌──────────────┐
                       │ GPT-5-Nano   │
                       │ (Reasoning)   │
                       └──────────────┘

🚀 Quick Start

Using Docker Compose (Recommended)

1. Environment Setup

   # Copy and configure environment
   cp .env.example .env
   
   # Edit .env with your API keys
   # OPENAI_API_KEY=your_openai_key
   # LIGHTRAG_API_KEY=your_lightrag_key (optional)
   

2. Deploy

   docker-compose up -d
   

3. Verify Deployment

   curl http://localhost:8000/health
   

Using Docker Directly

# Build the image
docker build -t cyberlegal-ai .

# Run the container
docker run -d \
  --name cyberlegal-ai \
  -p 8000:8000 \
  -e OPENAI_API_KEY=your_key \
  -v $(pwd)/rag_storage:/app/rag_storage \
  cyberlegal-ai

📡 API Usage

Base URL

http://localhost:8000

Endpoints

Chat with Assistant

curl -X POST "http://localhost:8000/chat" \
  -H "Content-Type: application/json" \
  -d '{
    "message": "What are the main obligations under GDPR?",
    "role": "client",
    "jurisdiction": "EU",
    "conversationHistory": []
  }'

Health Check

curl http://localhost:8000/health

API Info

curl http://localhost:8000/

📝 Request Format

{
  "message": "User's legal question",
  "role": "client" | "lawyer",
  "jurisdiction": "EU" | "France" | "Germany" | "Italy" | "Spain" | "Romania" | "Netherlands" | "Belgium",
  "conversationHistory": [
    {"role": "user|assistant", "content": "Previous message"}
  ]
}

📤 Response Format

{
  "response": "Detailed legal answer with references",
  "confidence": 0.85,
  "processing_time": 2.34,
  "references": ["gdpr_2022_2555.txt", "nis2_2022_2555.txt"],
  "timestamp": "2025-01-15T10:30:00Z",
  "error": null
}

🧠 Expertise Areas

• GDPR (General Data Protection Regulation)
• NIS2 (Network and Information Systems Directive 2)
• DORA (Digital Operational Resilience Act)
• CRA (Cyber Resilience Act)
• eIDAS 2.0 (Electronic Identification, Authentication and Trust Services)
• Romanian Civil Code provisions

🔄 Workflow

1. User Query → API receives request with role/jurisdiction context
2. LightRAG Retrieval → Searches legal documents for relevant information
3. LangGraph Processing → Orchestrates the workflow through nodes:
   • Query validation
   • LightRAG integration
   • Context enhancement with GPT-5-Nano
   • Response formatting
4. Enhanced Response → Returns structured answer with confidence score

🛠️ Development

Local Development

# Install dependencies
pip install -r requirements.txt

# Start LightRAG server (required)
lightrag-server --host 127.0.0.1 --port 9621

# Start the API
python agent_api.py

Environment Variables

OPENAI_API_KEY=your_openai_api_key
LIGHTRAG_API_KEY=your_lightrag_api_key
LIGHTRAG_HOST=127.0.0.1
LIGHTRAG_PORT=9621
API_PORT=8000

📁 Project Structure

CyberlegalAI/
├── agent_api.py          # FastAPI server
├── langraph_agent.py     # Main LangGraph workflow
├── agent_state.py        # State management
├── prompts.py           # System prompts
├── utils.py             # LightRAG integration
├── requirements.txt      # Python dependencies
├── Dockerfile          # Container configuration
├── docker-compose.yml   # Orchestration
├── rag_storage/        # LightRAG data persistence
└── .env               # Environment variables

🔧 Configuration

Port Management

• Port 8000: API (exposed externally)
• Port 9621: LightRAG (internal only, for security)

Security Features

• LightRAG server not exposed externally
• API key authentication support
• Non-root container execution
• Health checks and monitoring

📊 Monitoring

Health Checks

# Container health
docker ps

# Service health
curl http://localhost:8000/health

# Logs
docker logs cyberlegal-ai

Performance Metrics

The API returns:

• Processing time per request
• Confidence scores
• Referenced documents
• Error tracking

🚨 Error Handling

The API gracefully handles:

• LightRAG server unavailability
• OpenAI API errors
• Invalid request format
• Network timeouts

📚 API Examples

Client Role Example

{
  "message": "What should my small business do to comply with GDPR?",
  "role": "client",
  "jurisdiction": "France"
}

Lawyer Role Example

{
  "message": "Analyze the legal implications of NIS2 for financial institutions",
  "role": "lawyer",
  "jurisdiction": "EU"
}

Comparison Query

{
  "message": "Compare incident reporting requirements between NIS2 and DORA",
  "role": "client",
  "jurisdiction": "EU"
}

🤝 Integration Examples

Python Client

import requests

response = requests.post("http://localhost:8000/chat", json={
    "message": "What are GDPR penalties?",
    "role": "client",
    "jurisdiction": "EU",
    "conversationHistory": []
})

result = response.json()
print(result["response"])

JavaScript Client

const response = await fetch('http://localhost:8000/chat', {
  method: 'POST',
  headers: { 'Content-Type': 'application/json' },
  body: JSON.stringify({
    message: 'GDPR requirements',
    role: 'client',
    jurisdiction: 'EU',
    conversationHistory: []
  })
});

const result = await response.json();
console.log(result.response);

📋 Troubleshooting

Common Issues

1. LightRAG Connection Failed

   • Verify LightRAG server is running on port 9621
   • Check container logs: docker logs cyberlegal-ai

2. OpenAI API Errors

   • Verify OPENAI_API_KEY is set correctly
   • Check API key permissions and quota

3. Slow Responses

   • Monitor processing time in API response
   • Check LightRAG document indexing

Debug Mode

Enable debug logging:

docker-compose logs -f cyberlegal-api

📜 License

This project provides general legal information and is not a substitute for professional legal advice.

🔄 Updates

The system automatically:

• Retrieves latest regulatory documents
• Updates knowledge base through LightRAG
• Maintains conversation context
• Provides confidence scoring