mcp-google-cse

mcp-google-cse

21
cloud_platforms

Google Custom Search Engine MCP Server allows language models to perform searches using a Google custom search engine, returning structured search results. Designed as a Model Context Protocol server, it supports tool chaining and multiple installation options. It offers a free quota for 100 searches daily.

Overview

Google Custom Search Engine MCP Server is a Model Context Protocol server that provides search capabilities using a custom search engine. It allows language models to perform Google-like searches and return search results such as titles, links, and snippets. Users can combine it with other servers for content extraction and deep search functionalities. Key features include a free quota of 100 searches per day, tool chaining potential, and configuration options for different environments including UV and PIP installation.

Available Tools

  • google_search: Searches the custom search engine using the search term and returns a list of results.

Environment Variables

  • API_KEY: API key for the custom search engine.
  • ENGINE_ID: Engine ID for the custom search engine.
  • SERVICE_NAME: Name of the service.
  • COUNTRY_REGION: Restricts search results to a country.
  • GEOLOCATION: Sets geolocation of the search.
  • RESULT_LANGUAGE: Language of the search results.
  • RESULT_NUM: Number of results to return.

CSE Setup

Instructions to set up a Google custom search engine via Google console and programmable search engine page.

Installation

Using UV

Directly run mcp-google-cse with uvx.

Using PIP

Install via pip and run as a script.

Installing via Smithery

Install for Claude Desktop via Smithery CLI.

Configuration

Configure for Claude app

JSON configuration examples for running using different methods.