Gemini Media

stdio

Generate images (Nano Banana), video (Veo, Gemini Omni), speech and music (Lyria) with Google models

Getting Started

npx @getmcp/cli add io.github.mordor-forge/gemini-media-mcp

Requires Docker Engine. Community-contributed server. Review source before installing.

Claude Desktopguide →Vscodeguide →Cursorguide →Windsurfguide →

Required Environment Variables

GEMINI_API_KEY
Gemini API key from https://aistudio.google.com/apikey.

Configuration

Config file: ~/Library/Application Support/Claude/claude_desktop_config.json

JSON

{
  "mcpServers": {
    "io.github.mordor-forge/gemini-media-mcp": {
      "command": "docker",
      "args": [
        "{host_output_dir}:/output",
        "gemini-media-state:/state",
        "MEDIA_OUTPUT_DIR=/output",
        "run",
        "-i",
        "--rm",
        "-e",
        "GEMINI_API_KEY",
        "-e",
        "GEMINI_MEDIA_BUDGET_DAILY_USD",
        "ghcr.io/mordor-forge/gemini-media-mcp:1.0.0"
      ],
      "env": {
        "GEMINI_API_KEY": "",
        "GEMINI_MEDIA_BUDGET_DAILY_USD": ""
      }
    }
  }
}

Claude Desktop MCP documentation