Google Search MCP サーバー

What is Google Search MCP サーバー?

Google Search MCP サーバー is a Model Context Protocol (MCP) server that provides AI assistants with Google Search API capabilities. It uses Gradio and MCP to act as an interface between AI assistants and the Google Search API, enabling web search functionality.

How to use Google Search MCP サーバー?

Install dependencies with pip install -e ., set the GOOGLE_CSE_ID and GOOGLE_API_KEY environment variables in a .env file, then start the server with python app.py. The server launches a Gradio interface that functions as an MCP endpoint.

Key features of Google Search MCP サーバー

• Performs web searches using the Google Custom Search API
• Returns search results in JSON format
• Supports up to 20 results per query
• Integrates with AI assistants via the Model Context Protocol (MCP)
• Requires Python 3.12 or later

Use cases of Google Search MCP サーバー

• Enabling AI assistants to fetch real‑time web data
• Answering queries that require up‑to‑date or external information
• Fact‑checking or verifying claims during a conversation
• Integrating Google Search into custom AI workflows and chat interfaces

FAQ from Google Search MCP サーバー

What prerequisites are needed?

The server requires Python 3.12+, a Google Custom Search Engine ID (CSE ID), and a Google API key. Both credentials must be set in a .env file.

How many search results can I get per request?

You can request between 1 and 20 results; the default is 10. The num_results parameter controls this.

Are there limits on the search query?

The query string is limited to 400 characters (approximately 50 words). The perform_web_search function enforces this limit.

Where does the data come from?

All search results are fetched from the Google Custom Search API. No local caching or data storage is described in the README.

How does authentication work?

Authentication is handled via the Google Custom Search API key and CSE ID set as environment variables. No additional OAuth or user‑level authentication is mentioned.