Gemini MCP Server

Give Claude superpowers with Gemini's massive context window. Analyze entire codebases, search the web, process images, and brainstorm ideas—all through MCP.

graph LR
    A[Claude Code] --> B[Gemini MCP Server]
    B --> C[gemini]
    B --> S[web-search]
    B --> M[analyze-media]
    B --> SH[shell]
    B --> D[brainstorm]
    C --> H[Gemini CLI]
    S --> H
    M --> H
    SH --> H
    D --> H
    H --> E[Extensions]
    E --> W[Google Workspace]
    E --> X[Custom Extensions]

    style A fill:#FF6B35
    style B fill:#4A90E2
    style C fill:#00D4AA
    style S fill:#00D4AA
    style M fill:#00D4AA
    style SH fill:#00D4AA
    style D fill:#00D4AA
    style H fill:#4285F4
    style E fill:#FBBC05
    style W fill:#34A853
    style X fill:#EA4335

Quick Start

Prerequisites

1. Install Gemini CLI: npm install -g @google/gemini-cli
2. Authenticate: Run gemini and login with Google (free: 60 req/min, 1000 req/day)
3. Set default model (recommended): export GEMINI_MODEL="gemini-3-pro-preview"

Install

One-Click

Manual

# Claude Code
claude mcp add gemini-cli -- npx -y @tuannvm/gemini-mcp-server

Add to ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%/Claude/claude_desktop_config.json (Windows):

{
  "mcpServers": {
    "gemini-cli": {
      "command": "npx",
      "args": ["-y", "@tuannvm/gemini-mcp-server"]
    }
  }
}

What You Can Do

| Tool | What it does | |------|--------------| | gemini | Analyze files and codebases with Gemini's 1M+ token context | | web-search | Search the web with Google Search grounding | | analyze-media | Process images, PDFs, and screenshots | | shell | Generate and execute shell commands | | brainstorm | Creative ideation with structured methodologies |

Gemini CLI Extensions

Enable access to Gmail, Drive, Sheets, Docs, Calendar, and more through Gemini CLI extensions.

# Install the Google Workspace extension
gemini extensions install https://github.com/gemini-cli-extensions/workspace

Then use with yolo: true to auto-approve extension tool calls:

# Search your Google Drive
Use gemini with yolo=true to search my Google Drive for Q4 reports

# Read a Google Sheet
Use gemini with yolo=true to read the first sheet from "Budget 2025"

# Access Gmail
Use gemini with yolo=true to summarize my last 10 unread emails

Note: Any Gemini CLI extension can be used. Pass the full GitHub URL or local path to gemini extensions install.

Examples

# Analyze an entire codebase
Use gemini to analyze @. and provide an architecture overview

# Compare files
Use gemini to compare @src/old.ts and @src/new.ts

# Search the web
Use web-search to find the latest React 19 features

# Analyze a screenshot
Use analyze-media with filePath "@error.png" and prompt "what's this error?"

# Brainstorm ideas
Use brainstorm with methodology "design-thinking" to improve user onboarding

Documentation

• API Reference — Full tool parameters and options
• Gemini CLI Integration — Setup, models, and troubleshooting
• Documentation Site

Development

npm install    # Install dependencies
npm run dev    # Development mode
npm run build  # Build
npm test       # Run tests

Acknowledgments

Inspired by gemini-mcp-tool.

Related Projects

• codex-mcp-server — MCP server for OpenAI Codex CLI with AI-powered code analysis and review

License

MIT — This is an unofficial tool, not affiliated with Google.