CrewAI MCP Adapter

A Python library extending CrewAI’s adapter ecosystem with Model Context Protocol (MCP) integration support and comprehensive tooling for custom agent and tool development.

Features

• 🔌 Native CrewAI integration and adapter patterns
• 🛠️ MCP protocol support for tool integration
• 🧩 Easy-to-use interface for extending and creating new adapters
• 📝 Type-safe implementation with Pydantic
• 🔍 JSON Schema validation for tool parameters
• 🚀 Async/await support
• 📊 Detailed execution metadata

Installation

You can install the package directly from PyPI:

pip install crewai-adapters

Or install from source:

pip install git+https://github.com/dshivendra/crewai_mcp_adapter.git

Quick Start

from crewai import Agent, Task
from crewai_adapters import CrewAIAdapterClient
from crewai_adapters.types import AdapterConfig

async def main():
    async with CrewAIAdapterClient() as client:
        # Connect to MCP server
        await client.connect_to_mcp_server(
            "math",
            command="python",
            args=["math_server.py"]
        )

        # Create agent with tools
        agent = Agent(
            name="Calculator",
            goal="Perform calculations",
            tools=client.get_tools()
        )

        # Execute task
        task = Task(
            description="Calculate (3 + 5) × 12",
            agent=agent
        )
        result = await task.execute()
        print(f"Result: {result}")

if __name__ == "__main__":
    import asyncio
    asyncio.run(main())

Documentation

For detailed documentation, see:

• Getting Started Guide
• API Reference
• Examples

Development

Prerequisites

• Python 3.11 or higher
• crewai package
• pydantic package
• mcp package

Install Development Dependencies

pip install -e ".[test,docs]"

Running Tests

pytest tests/ -v

Publishing

To publish a new version to PyPI:

1. Update version in pyproject.toml
2. Build the package:

   python -m build
   

3. Upload to PyPI:

   python -m twine upload dist/*
   

License

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

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.