Repolode
Submit
MiniMax-AI

MiniMax-AI/MiniMax-MCP

Official MCP server that lets MCP clients like Claude Desktop and Cursor call MiniMax APIs for text-to-speech, voice cloning, image and video generation.

View on GitHubWebsite
Stars
1,594
Forks
285
Open issues
41
Last commit
2 months ago

What is it?

MiniMax-MCP is the official Model Context Protocol server for MiniMax's generation APIs. It wraps them as MCP tools so any compatible client, such as Claude Desktop, Cursor, Windsurf or OpenAI Agents, can generate speech from text, clone a voice from audio files, design a voice from a text prompt, and create images or videos from prompts.

The server runs over two transports: stdio for local use, where it reads local files and writes output to a directory, and SSE for deployment on a server, where URL inputs are recommended. It is installed and launched with uvx, and it needs a MiniMax API key whose region matches the API host (global or mainland China).

Video generation is asynchronous: a generate_video call starts a task, and query_video_generation retrieves the result. The project is a Python package released under the MIT license.

Key features

  • 01Text-to-speech with selectable voices and a list_voices tool to browse them.
  • 02Voice cloning from provided audio files and voice design from a text prompt.
  • 03Text-to-image and text-to-video generation, including image-to-video inputs.
  • 04Supports stdio for local clients and SSE for cloud deployment.
  • 05Works with MCP clients such as Claude Desktop and Cursor via a simple JSON config.

Use cases

  • Adding speech synthesis or voice cloning to an AI assistant running in Claude Desktop or Cursor.
  • Generating images or short videos from prompts inside an agent workflow.
  • Hosting a shared MCP server over SSE so a team's agents can call MiniMax generation tools.

Topics

#mcp#mcp-server#text-to-speech#voice-cloning#image-generation#video-generation#ai-agent#image-to-video#mcp-tools#text-to-image

FAQ

What is MiniMax-MCP?
Official MCP server that lets MCP clients like Claude Desktop and Cursor call MiniMax APIs for text-to-speech, voice cloning, image and video generation. MiniMax-MCP is the official Model Context Protocol server for MiniMax's generation APIs. It wraps them as MCP tools so any compatible client, such as Claude Desktop, Cursor, Windsurf or OpenAI Agents, can generate speech from text, clone a voice from audio files, design a voice from a text prompt, and create images or videos from prompts.
Is MiniMax-MCP free to use? What license does it have?
MiniMax-MCP is open source under the MIT license. Check the license terms before using it in commercial products.
Is MiniMax-MCP still maintained?
Updated within 6 months. The last commit to MiniMax-MCP was on Aug 20, 2026.
What can I use MiniMax-MCP for?
Adding speech synthesis or voice cloning to an AI assistant running in Claude Desktop or Cursor. Generating images or short videos from prompts inside an agent workflow. Hosting a shared MCP server over SSE so a team's agents can call MiniMax generation tools.
Where can I get MiniMax-MCP?
The source code and installation instructions are on GitHub: https://github.com/MiniMax-AI/MiniMax-MCP

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects