google-knowledge-graph

This MCP server provides a bridge between Claude (or any MCP client) and Google's Knowledge Graph Search API. It exposes two main capabilities: search_knowledge_graph for querying entities by name or description, and lookup_knowledge_graph_entities for resolving machine IDs (MIDs) to full entity data. When invoked by an MCP client, the server returns structured JSON containing entity names, types, descriptions, Wikipedia URLs, images, and a relevance score. Use the search capability to discover entities related to a query, and use the lookup capability to resolve specific MIDs you already have. The server handles API requests to the Google Knowledge Graph API and returns results in a consistent MCP-friendly JSON format that downstream tools can easily parse and display.

To use, configure Claude Desktop (or your MCP host) with the provided server entry. You’ll supply your Google Knowledge Graph API key via the GOOGLE_KNOWLEDGE_GRAPH_API_KEY environment variable. Once configured and the MCP host is restarted, you can issue queries like searching for the term of interest or lookup by MID to get detailed entity records, including names, types, descriptions, and URLs, ready for integration into research workflows or knowledge apps.

Prerequisites:

• Node.js (LTS) and npm installed on your machine
• Git installed to clone repositories if you prefer running from source

Option 1: NPX (Easiest)

• No installation required beyond having Node.js/npm available. Use the NPX command provided in the config example to run directly from the npm registry.

Option 2: Local Install (from source)

1. Install Git and Node.js/npm
2. Clone the repository and install dependencies

git clone https://github.com/houtini/google-knowledge-graph-mcp.git
cd google-knowledge-graph-mcp
npm install
npm run build

3. Run or reference the built output from dist/index.js in your MCP config (see example below)

Configuring Claude Desktop after local install (example):

{
  "mcpServers": {
    "google-knowledge-graph": {
      "command": "node",
      "args": ["C:\\path\\to\\google-knowledge-graph-mcp\\dist\\index.js"],
      "env": {
        "GOOGLE_KNOWLEDGE_GRAPH_API_KEY": "your-api-key-here"
      }
    }
  }
}

Notes:

• Replace C:\path\to\... with your actual path. Use double backslashes on Windows as shown.
• The build step outputs CommonJS modules in dist/ suitable for MCP hosts.

Prerequisites recap: Node.js installed, optionally Git for source checkout, and an API key for Google Knowledge Graph.

Tips and troubleshooting:

• Ensure the API key environment variable GOOGLE_KNOWLEDGE_GRAPH_API_KEY is correctly set in your MCP host configuration and that Claude Desktop (or your MCP client) has access to it.
• If you encounter API key errors, verify the key is valid, has Knowledge Graph Search API enabled, and that it is restricted appropriately only if desired.
• When using NPX, you don’t manage a local install, but with a local install you’ll need to rebuild after code changes.
• If you see JSON parsing or MCP compatibility errors, confirm the MCP host is using a CommonJS module (the default for this server).
• Windows users: remember to escape backslashes in file paths (e.g., C:\MCP\path\to\dist\index.js).
• The Knowledge Graph Search API has usage quotas; monitor your token usage accordingly.