Gemini MCP Server

A Model Context Protocol (MCP) server for Google Gemini API integration with Claude Code. Provides text generation, image analysis, creative brainstorming, multi-turn chat, and AI image generation.

Features

Setup

1. Clone the repository

git clone https://github.com/jeffemmett/gemini-mcp.git ~/.claude/mcp-servers/gemini
cd ~/.claude/mcp-servers/gemini
npm install

2. Get a Gemini API Key

1. Go to Google AI Studio

2. Create a new API key

3. Copy the key

3. Set your API key

Add to your shell profile (~/.bashrc, ~/.zshrc, or ~/.config/fish/config.fish):

# Bash/Zsh
export GEMINI_API_KEY="your_api_key_here"

# Fish
set -gx GEMINI_API_KEY "your_api_key_here"

Then reload your shell or run source ~/.bashrc.

4. Configure Claude Code

Add to your project's .mcp.json or global MCP config:

{
  "mcpServers": {
    "gemini": {
      "command": "node",
      "args": ["/path/to/.claude/mcp-servers/gemini/index.js"],
      "env": {
        "GEMINI_API_KEY": "${GEMINI_API_KEY}"
      }
    }
  }
}

Replace /path/to/ with your actual home directory path (e.g., /home/username/ or /Users/username/).

Optional Configuration

Output Directory

Generated images and saved outputs go to ~/Documents/gemini-output/ by default. Override with:

export GEMINI_OUTPUT_DIR="/your/preferred/path"

Usage Examples

Once configured, the tools are available in Claude Code:

Text Generation:

Use gemini_generate to write a haiku about programming

Image Analysis:

Use gemini_analyze_image to describe /path/to/image.jpg

Creative Brainstorming:

Use gemini_brainstorm for a zine about urban gardening with punk style

Image Generation:

Use gemini_generate_image to create a collage-style illustration of a mushroom forest

Models Used

• Text: gemini-1.5-flash (default), gemini-1.5-pro, gemini-2.0-flash-exp

• Images: gemini-2.0-flash-exp-image-generation

License

MIT