Gemini MCP Server for Claude Desktop

Name: Gemini MCP Server for Claude Desktop
Author: Garblesnarff

@Garblesnarff

About Gemini MCP Server for Claude Desktop

A Model Context Protocol (MCP) server that enables Claude Desktop to interact with Google's Gemini AI - featuring 7 tools with Smart Tool Intelligence that learns from your usage patterns

Basic information

Config

Add this server to your MCP-compatible client using the configuration below.

{
  "mcpServers": {
    "gemini-mcp-server": {
      "command": "node",
      "args": [
        "test-tool-intelligence-full.js"
      ]
    }
  }
}

Tools

No tools detected

We auto-extract tools from the README. The maintainer can list them under a ## Tools heading to populate this section.

Overview

What is Gemini MCP Server for Claude Desktop?

Gemini MCP Server for Claude Desktop is an MCP server that provides 7 AI-powered tools via the Gemini API, featuring a Smart Tool Intelligence system that automatically learns user preferences and enhances prompts over time. It is designed for integration with Claude Desktop.

How to use Gemini MCP Server for Claude Desktop?

Clone the repository, run npm install, set your Gemini API key in a .env file, then start the server with npm start. Add a gemini entry to your Claude Desktop configuration (claude_desktop_config.json) pointing to the server script and including your API key in the environment variables.

Key features of Gemini MCP Server for Claude Desktop

7 AI tools: image generation, editing, chat, audio transcription, code execution, video analysis, image analysis
Smart Tool Intelligence: self-learning context detection and prompt enhancement
Persistent memory of preferences across sessions with automatic migration
Verbatim audio transcription mode capturing filler words and emotions
Secure sandboxed Python code execution
Supports multiple file types with defined size limits (images/audio 20MB, video 100MB)

Use cases of Gemini MCP Server for Claude Desktop

Generate images from text descriptions using Gemini 2.0 Flash
Edit existing images with natural language instructions
Transcribe audio files with optional verbatim mode for exact word-for-word output
Execute Python code in a sandboxed environment for data analysis or automation
Analyze video content for summaries, transcripts, and object detection

FAQ from Gemini MCP Server for Claude Desktop

How does the Smart Tool Intelligence system work?

It automatically detects context (e.g., coding, consciousness research, debugging), enhances prompts with relevant instructions, stores successful interaction patterns, and adapts over time.

Where are learned preferences stored?

Preferences are stored internally at ./data/tool-preferences.json with automatic migration from external storage.

What are the file size limits?

Images are limited to 20MB, audio to 20MB, and video to 100MB. Supported formats are listed for each tool.

What context types does the system recognize?

It recognizes consciousness (academic rigor), code (practical examples), debugging (root cause analysis), general (comprehensive responses), and verbatim (exact transcription for audio).