Model Context Protocol (MCP) is a popular open standard from Anthropic that provides AI models (usually LLMs) with context: resources, tools and more.

Agents can be viewed as MCP resources and further invoked as MCP tools. Such look at ACP agents allows MCP clients to discover and run ACP agents at a cost of reduced interface compared to full ACP.

​
Adapter

ACP-MCP Adapter is a production-ready MCP server that can be deployed in front of any ACP server to make the aforementioned bridging. It exposes agents as resources and defines tools to run them.

To run the adapter, there are currently two options:

  1. Use PyPI package
uvx acp-mcp <acp-server-url> # e.g. http://localhost:8000
  1. Use Docker image
docker run -i --rm ghcr.io/i-am-bee/acp-mcp <acp-server-url> # e.g. http://host.docker.internal:8000

The adapter currently supports stdio transport. Follow the documentation of your client to integrate one of the commands above (e.g. Claude tutorial).