MCP Google Custom Search Server

What is MCP Google Custom Search Server?

A Model Context Protocol (MCP) server that enables Language Learning Models (LLMs) to perform web searches through Google's Custom Search API. It provides a standardized, type-safe interface compatible with Claude Desktop and other MCP clients.

How to use MCP Google Custom Search Server?

Install dependencies with npm install, create a .env file with your GOOGLE_API_KEY and GOOGLE_SEARCH_ENGINE_ID, then build (npm run build) and start (npm start) the server. For Claude Desktop integration, add a JSON configuration entry pointing to the built server and supplying the same environment variables.

Key features of MCP Google Custom Search Server

• Seamless integration with Google Custom Search API
• MCP-compliant server implementation with TypeScript
• Configurable search results (up to 10 per query)
• Formatted output with titles, URLs, and descriptions
• Input validation using Zod
• Error handling and validation

Use cases of MCP Google Custom Search Server

• Providing web search results to an LLM for answering queries
• Integrating real-time web data into conversational AI assistants
• Enabling dynamic fact-checking and research via MCP clients

FAQ from MCP Google Custom Search Server

What prerequisites are needed to run the server?

You need a Google Cloud Project with the Custom Search API enabled, an API key, and a Custom Search Engine ID. Locally, Node.js v18 or higher and npm are required.

How do I configure the server for Claude Desktop?

Add a JSON block in the Claude Desktop config file (typically ~/Library/Application Support/Claude/claude_desktop_config.json) under mcpServers with the command node, the absolute path to the built index.js, and the required GOOGLE_API_KEY and GOOGLE_SEARCH_ENGINE_ID environment variables.

How many results can the search tool return per query?

The optional numResults parameter defaults to 5 and can be set up to a maximum of 10 results per search.

What runtime and dependencies does the server require?

It runs on Node.js v18+ with npm, and depends on the Google Custom Search API and the Model Context Protocol libraries.

Does the server handle invalid input or errors?

Yes, it includes input validation via Zod and error handling for API failures and malformed requests.