home / mcp / google knowledge graph search mcp

Google Knowledge Graph Search MCP

Provides access to Google’s Knowledge Graph Search API for querying real-world entities and retrieving structured data via MCP.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "houtini-ai-google-knowledge-graph-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "@houtini/google-knowledge-graph-mcp"
      ],
      "env": {
        "GOOGLE_KNOWLEDGE_GRAPH_API_KEY": "YOUR_API_KEY"
      }
    }
  }
}

This MCP server lets Claude (or any MCP client) access Google's free Knowledge Graph Search API to retrieve structured information about real-world entities. It enables you to search by query or look up entities by their machine IDs, returning rich data such as names, types, descriptions, images, URLs, and relevance scores for use in your workflows.

How to use

You install this server and connect it to your MCP client. Use the provided tools to search for entities by name or description or to resolve entities by their machine IDs. The server returns structured JSON containing entity names, types, descriptions, official images, URLs, and a relevance score, which you can feed directly into your research or verification workflows.

How to install

Prerequisites: Node.js and npm installed on your system. An API key for the Knowledge Graph Search API is required.

{
  "mcpServers": {
    "google-knowledge-graph": {
      "command": "npx",
      "args": ["-y", "@houtini/google-knowledge-graph-mcp"],
      "env": {
        "GOOGLE_KNOWLEDGE_GRAPH_API_KEY": "your-api-key-here"
      }
    }
  }
}

Alternatively, you can run from source with a local install.

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

After building, configure your MCP client to start the local server using the built index. Replace the path with your actual build path.

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

Additional sections

Configuration details and troubleshooting help ensure you can get the MCP server running smoothly and securely.

Troubleshooting

If the server does not appear in your MCP client, verify JSON syntax, ensure environment variables are correctly spelled, and restart your client after configuration changes. If you encounter an API key error, confirm the key is set as GOOGLE_KNOWLEDGE_GRAPH_API_KEY and that the key has access to the Knowledge Graph Search API.

Building From Source

If you built the project from source, you can start the runtime from the generated dist/index.js using the runtime command shown in the local install snippet.

Technical Details

This MCP server communicates with Google's Knowledge Graph Search API using an API key. The server returns structured data including entity names, types, descriptions, images, URLs, and a relevance score for each result.

Available tools

search_knowledge_graph

Search entities by name or description and receive structured results including types, descriptions, Wikipedia URLs, and relevance scores.

lookup_knowledge_graph_entities

Lookup entities directly by their machine IDs (MIDs) to support entity resolution workflows.