Revit Api Mcp

Tools

This repository provides four MCP tools for working with Revit API documentation:

• search-docs - Search Revit API documentation to find entities matching your query. Returns entity names, descriptions, namespaces, types, and URL slugs for further exploration (but not the documentation itself).

• retrieve-doc - Retrieve a single Revit API documentation page using its URL slug. Use this after getting a URL slug from a search operation.

• retrieve-docs - Get full documentation content for multiple Revit API entities based on a search query. Useful when you need complete documentation content, not just search results.

• search-library - Search a comprehensive library of Revit API learning resources including blog posts, code examples, PDFs, and practical guides. Powered by OpenAI's vector store for semantic search.

Setup

Download the executable for your OS from Releases. Or you can clone and build from source using deno task compile. Add this executable somewhere in your file system that makes sense. Good practice for Windows is <username>/bin/, but it can be anywhere.

Run path\to\executable -h in your terminal to see the help menu. FYI: This executable is useless to run in the terminal besides for the help menu and for testing the command you will be adding to your mcp config.

Abbreviated Help Menu

To use the MCP server, add the executable to your mcp config for whatever MCP client and OS your using.

{
  "mcpServers": {
    "revit-api-docs (macos-arm64)": {
      "command": "path/to/Rvt_Docs_MCP-macos-arm64"
    },
    "revit-api-docs (macos-x64)": {
      "command": "path/to/Rvt_Docs_MCP-macos-x64"
    },
    "revit-api-docs (windows)": {
      "command": "path\\to\\Rvt_Docs_MCP-windows.exe"
    },
    "revit-api-docs (with search-library enabled)": {
      "command": "path\\to\\Rvt_Docs_MCP-windows.exe -k 'sk-xxx' -v 'vs_xxx'"
    }
  }
}

To enable search-library, you must first follow the steps described in Rvt_Docs_Tbc_Embedder. After doing so, run the executable with the -k (OpenAI API key) and -v (OpenAI vector store ID) flags as seen above. To verify that this runs properly, run the command on its own in your terminal. The ouput is informative, though as mentioned above its meaningless otherwise.