gemini-mcp-power

gemini-mcp-power

MCP server for Google Gemini AI with custom proxy endpoint support.

Features

• Text Generation - Generate text using Gemini models
• Image Generation - Create images from text prompts
• Speech Generation (TTS) - Convert text to speech with multiple voices
• Audio Transcription - Transcribe audio files to text
• Video Generation - Generate videos from text prompts (async)

Installation

npm install -g gemini-mcp-power

Configuration

Add to your MCP config (e.g., .kiro/settings/mcp.json):

{
  "mcpServers": {
    "gemini": {
      "command": "npx",
      "args": ["-y", "gemini-mcp-power"],
      "env": {
        "GEMINI_API_KEY": "your-api-key",
        "GEMINI_BASE_URL": "https://your-proxy-endpoint.com"
      }
    }
  }
}

Environment Variables

• GEMINI_API_KEY (required) - Your Gemini API key
• GEMINI_BASE_URL (optional) - Custom proxy endpoint URL
• GEMINI_MODEL_TEXT (optional) - Text generation model (default: gemini-3-flash)
• GEMINI_MODEL_IMAGE (optional) - Image generation model (default: gemini-3-pro-image-preview)
• GEMINI_MODEL_SPEECH (optional) - TTS model (default: gemini-2.5-flash-preview-tts)
• GEMINI_MODEL_VIDEO (optional) - Video generation model (default: veo-3.1-generate-preview)

Available Tools

Tool	Description
gemini_text	Generate text
gemini_generate_image	Generate images
gemini_generate_speech	Text-to-speech
gemini_transcribe_audio	Audio transcription
gemini_generate_video	Start video generation
gemini_check_video	Check video status and download

Usage Examples

Text Generation

Generate a product description for a smartwatch

Image Generation

Generate an image of a sunset beach, save to /path/to/image.png

Speech Generation

Supported voices: Puck, Charon, Kore, Fenrir, Aoede

Convert text to speech, save to /path/to/audio.wav

Video Generation

Video generation is async - use gemini_generate_video to start, then gemini_check_video to download when ready.

License

MIT