MiniMax MCP 🔌 MCP server Freemium

What it is

MiniMax MCP is the official Model Context Protocol server from MiniMax, the Chinese AI lab behind the Hailuo video models and Speech voices. It is an open-source Python server that runs locally and calls the MiniMax API, so an MCP client such as Claude Desktop, Cursor or Windsurf can produce speech, cloned voices, video clips and images from a chat.

Key features

• Text-to-speech with a chosen MiniMax voice
• Voice cloning from audio files you provide
• Voice design: create a new voice from a text description
• Video generation from a prompt, including MiniMax-Hailuo-02
• Image generation from a prompt
• Query tool to poll long video jobs until they finish

How to install

Install uv, then add a server entry that runs uvx minimax-mcp with three environment variables: MINIMAX_API_KEY, MINIMAX_API_HOST (https://api.minimax.io for the global platform, https://api.minimaxi.com for mainland China) and MINIMAX_MCP_BASE_PATH for output files. The README warns that the key must match the host, otherwise you get authentication errors.

Pricing and rights

The server is free and MIT-licensed; generation is billed by the MiniMax API. As of September 2026 MiniMax sells audio subscriptions from $5 per month (Starter, 100,000 audio points) up to $999, with pay-as-you-go pricing for video and image. Check MiniMax's terms for commercial use of outputs; the README does not cover rights.

Who it is for

Developers and creators who want Hailuo video and MiniMax voices inside an agent workflow, especially if they already hold a MiniMax API key.

Verdict

One server covers voice, video and image, which is rare, and it comes from the model maker. Setup is fiddlier than hosted servers (uv, region-matched keys, local output folder) and the video tools are asynchronous, so expect to poll for results.