MCP Servers

Model Context Protocol (MCP) servers are external applications that expose tools, data, and functionality to the Large Language Models (LLMs) used in ChatFrame. They are the key to extending the AI's capabilities with real-time context.

What is an MCP Server?

An MCP server is a program that communicates with ChatFrame (the MCP host/client) using the JSON-RPC 2.0 protocol [1]. It defines a set of functions (tools) that the LLM can call during a conversation.

How to Use MCP Servers in ChatFrame

  1. Install the Server: Download and run the MCP server application you wish to use (e.g., a server for querying a PostgreSQL database, a server for web search, or a custom internal API server).
  2. Configure in ChatFrame: In ChatFrame's settings, you will configure the connection details for the running MCP server (e.g., its local network address and port).
  3. Activate in Chat: In the chat interface, you can select the active MCP server. All the tools exposed by that server will then be available for the LLM to use.

Examples of MCP Server Use Cases

Use CaseDescriptionExample Server
Database QueryAllows the LLM to write and execute SQL queries against a database (e.g., PostgreSQL) to retrieve real-time data.postgres-example.md
Web SearchProvides the LLM with up-to-date information from the internet, overcoming its knowledge cutoff.N/A (TBD)
Internal ToolsExposes custom internal APIs or scripts to the LLM, enabling it to automate complex business workflows.Custom-built server

Runtime Requirements

Remember that you must have the necessary runtime installed for the specific MCP server you choose to run. ChatFrame does not bundle runtimes like Node.js or Python [2].

  1. Model Context Protocol. Specification. https://modelcontextprotocol.io/specification/2025-03-26
  2. ChatFrame. Official Website. https://www.chatframe.co/