AI Agent Management System

A scalable system for managing, monitoring, and orchestrating AI agents built with FastAPI, LangChain, and SQLite.

🚀 Features

• Agent Management

  • Create and configure AI agents
  • Start/stop agent execution
  • Monitor agent status and performance
  • Real-time updates via WebSocket

• Task Execution

  • Assign tasks to agents
  • Track task progress
  • Store execution history
  • Handle errors and retries

• Resource Management

  • Database persistence
  • Memory management
  • Tool integration
  • State tracking

🛠️ Technical Stack

• Backend: FastAPI
• Database: SQLite
• AI Framework: LangChain
• LLM Integration: OpenAI
• API Documentation: Swagger/OpenAPI

📋 Prerequisites

• Python 3.9+
• OpenAI API key
• Virtual environment management

🔧 Installation

1. Clone the repository:

git clone https://github.com/yourusername/ai-agent-manager.git
cd ai-agent-manager

2. Create and activate virtual environment:

# Create virtual environment
python3 -m venv venv

# Activate (Mac/Linux)
source venv/bin/activate
# Activate (Windows)
# venv\Scripts\activate

3. Install dependencies:

pip install -r requirements.txt

4. Create .env file:

# .env
OPENAI_API_KEY=your_api_key_here
DATABASE_URL=sqlite:///agents.db
API_HOST=127.0.0.1
API_PORT=8000

🚀 Running the Application

1. Start the server:

python run.py

2. Access the API documentation:

http://127.0.0.1:8000/docs

📖 API Documentation

Endpoints

Agents

• POST /api/v1/agents/ - Create a new agent
• POST /api/v1/agents/{agent_id}/start - Start an agent
• POST /api/v1/agents/{agent_id}/stop - Stop an agent
• GET /api/v1/agents/{agent_id} - Get agent status

Tasks

• POST /api/v1/agents/{agent_id}/tasks - Execute a task

WebSocket

• WS /api/v1/ws/agents/{agent_id} - Real-time agent updates

Example Usage

# Create an agent
POST /api/v1/agents/
{
    "name": "research_agent",
    "config": {
        "model_name": "gpt-3.5-turbo",
        "temperature": 0.7
    }
}

# Execute a task
POST /api/v1/agents/{agent_id}/tasks
{
    "type": "research",
    "input": "What are the latest developments in AI?"
}

🗄️ Project Structure

ai-agent-manager/
├── src/
│   ├── api/           # API endpoints and routing
│   ├── config/        # Configuration and settings
│   ├── core/          # Core business logic
│   └── database/      # Database models and setup
├── tests/             # Test files
├── .env              # Environment variables
├── requirements.txt   # Python dependencies
└── run.py            # Application entry point

🧪 Running Tests

# Run all tests
pytest

# Run specific test file
pytest tests/test_agent_manager.py

# Run with coverage
pytest --cov=src tests/

🔒 Security

• All endpoints require authentication (TODO)
• API keys are managed via environment variables
• Database connections are properly managed
• Input validation on all endpoints

🤝 Contributing

1. Fork the repository
2. Create a feature branch
3. Commit your changes
4. Push to the branch
5. Create a Pull Request

📝 License

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

✨ Future Enhancements

• Add authentication and authorization
• Implement rate limiting
• Add more agent tools and capabilities
• Enhance monitoring and analytics
• Add support for multiple LLM providers
• Implement agent collaboration features

🐛 Known Issues

• Currently only supports single-user mode
• Limited to OpenAI's API
• Basic error handling

📞 Support

For support, please open an issue in the GitHub repository or contact the maintainers.

---

Made with ❤️ by [Your Name]