MiniMax MCP Server (Official)

Add to your client

Copy the config for your MCP client and paste it into its config file.

{
  "mcpServers": {
    "minimax-mcp-server-official": {
      "command": "uvx",
      "args": [
        "minimax-mcp",
        "-y"
      ],
      "env": {
        "MINIMAX_API_KEY": "<your-api-key>",
        "MINIMAX_API_HOST": "https://api.minimax.io",
        "MINIMAX_MCP_BASE_PATH": "/absolute/path/to/output/dir"
      }
    }
  }
}

Before you start

• Python with the uv package manager installed (the server is launched via uvx/uv)
• A MiniMax API key from your account at the MiniMax platform (platform key page, region-specific)
• The correct regional API host: https://api.minimax.io (global) or https://api.minimaxi.com (mainland China) — host and key must match the same region
• A writable local output directory for generated audio/image/video files
• An MCP-capable client (Claude Desktop, Cursor, Windsurf, etc.)

About MiniMax MCP Server (Official)

Tools & capabilities (8)

When to use it

• Use it when you need to turn scripts or articles into natural-sounding speech directly from your agent
• Use it when you want to clone or design a custom voice for narration, characters, or branded audio
• Use it when generating background music or jingles from a prompt for video or app projects
• Use it when you need on-demand image generation inside an MCP workflow
• Use it when producing short videos from text and need to poll the job until it completes

Quick setup

1. 1Install Python and the uv package manager
2. 2Sign in to MiniMax and create an API key, noting your region (global vs mainland)
3. 3Add the MiniMax server to your MCP client config (e.g. Claude Desktop: Settings > Developer > Edit Config), passing the API key, matching API host, and an output directory
4. 4Restart the MCP client so it picks up the new server
5. 5Verify by calling list_voices, then try a text_to_audio request

MiniMax MCP Server (Official) FAQ