This project provides a streamlined way to convert agents and orchestrators from popular frameworks like CrewAI and LangGraph into standardized MCP (Model Context Protocol) servers. It targets developers building AI agents who need to integrate them with clients like Cursor or Claude Desktop, enabling easier interoperability and deployment.

How It Works

The tool generates a run_mcp.py file that acts as a bridge. It uses adapter functions (e.g., create_crewai_adapter) to wrap existing agent logic. These adapters handle the conversion of agent inputs and outputs to the MCP format, allowing the agent to be served via STDIO or SSE transports. This approach abstracts away the complexities of different agent framework APIs and MCP communication protocols.

Quick Start & Requirements

• Install via pip: pip install naptha-automcp or uv add naptha-automcp.
• Clone and install from source: git clone https://github.com/NapthaAI/automcp.git && cd automcp && uv venv && source .venv/bin/activate && pip install -e .
• Requires Python.
• See Examples for framework-specific setup.

Highlighted Details

• Supports CrewAI, LangGraph, Llama Index, OpenAI Agents SDK, and Pydantic AI.
• Generates MCP server files via CLI (automcp init -f <FRAMEWORK>).
• Supports both STDIO and SSE transport modes for server communication.
• Includes detailed examples for each supported framework.

Maintenance & Community

• Developed by NapthaAI.
• Deployment to Naptha's MCPaaS platform is supported.
• Instructions for creating new adapters are provided.

Licensing & Compatibility

• The repository does not explicitly state a license in the README.

Limitations & Caveats

• For STDIO transport, print statements in agent code can corrupt the communication protocol.
• Users must ensure correct absolute paths are used when configuring clients like Cursor.
• The README does not specify a license, which may impact commercial use or closed-source integration.