Agent
A local MCP server integrating Google APIs with MCP services.
Install & Use
Copy this prompt and send it to your AI assistant (Claude / Cursor / TRAE / Codex / WorkBuddy etc.) to auto-install:
Help me install this AI Skill: Agent. It is used for: A local MCP server integrating Google APIs with MCP services. Details: https://321skill.com/en/skills/agent-agent-mcp/
This Skill addresses the inability to directly call Google APIs—such as Search and Knowledge Graph—for complex queries in local development environments. Many AI applications require real-time access to external information, but traditional approaches involve manually wrapping API requests, handling authentication, and managing rate limits, which reduces development efficiency. With this MCP server, you can seamlessly integrate Google APIs into the MCP protocol, enabling AI assistants to perform tasks like search and knowledge queries directly through natural language instructions.
Usage is straightforward: simply deploy the MCP server locally, configure your Google API key, and connect it within an MCP-compatible AI client (e.g., Claude Desktop, Cursor). You can then initiate queries via natural language commands, such as "Help me search for the latest AI papers." The MCP server will automatically call the Google Custom Search API and return structured results.
It is particularly suitable for developers who frequently need real-time external information, especially those building LLM-based agent applications or teams/individuals requiring internet search capabilities. It also works well in local development environments, eliminating reliance on third-party cloud services and offering greater data privacy control.
We recommend using this tool when regularly querying the latest materials, technical documentation, or market trends. Note that it primarily depends on Google API quota limits; high-frequency usage may require applying for paid tiers. Additionally, local deployment requires a Node.js environment, and it is advisable to keep the server running continuously to support real-time queries.
Key Features
Compared to official MCP servers (e.g., @modelcontextprotocol/server-google-search), this Skill offers more flexible local configuration, allowing customization of search parameters and caching strategies, and can easily extend to other Google APIs (such as Knowledge Graph and Custom Search), whereas the official package only supports a single search interface.
Limitations
Requires Node.js 18+ environment and a Google API key (must apply for Custom Search API or Knowledge Graph API). Does not support offline use, and network requests are subject to Google service availability.
FAQ
How do I obtain a Google API key?
Go to the Google Cloud Console to create a project, enable the Custom Search API or Knowledge Graph API, then generate an API key on the Credentials page and enter it in the configuration file.
Which AI clients does this MCP server support?
It supports all clients that adhere to the MCP protocol specification, such as Claude Desktop, Cursor, and Continue. Simply specify the MCP server's address and port in the client configuration.