Skip to main content
The MCP Agents category contains 13 projects that demonstrate how to connect LLM agents to external tools and services using the Model Context Protocol (MCP). MCP provides a standardized way for agents to discover and invoke external functionality, such as querying databases, reading GitHub repositories, searching documentation, and interacting with productivity tools.

MCP Agent Projects

Running an MCP Agent

Most MCP projects use a client-server architecture. The agent (client) connects to an MCP server that exposes tools the agent can call.
1

Navigate into the project

2

Configure environment keys

MCP projects usually require keys for both the LLM provider and the external service (e.g., GitHub token, Couchbase credentials, Exa API).
3

Install and run

Some projects also require starting the MCP server separately. Check the README for those instructions.
If you are new to MCP, start with the MCP Starter or Talk to your Docs projects. They clearly demonstrate how an agent discovers and calls tools through the MCP layer.