An MCP server brings media generation into AI coding tools

chutes-media-mcp is an MCP server that lets coding tools create images, videos, music, and speech without leaving the editor. It connects to Chutes media models, so a developer can describe the asset they need and have the result saved inside the project. The tool fetches each model’s live , which means it is not locked to one fixed model setup and can work across image, video, music, and speech models in the catalog.

The server handles sign-in, slow first starts, automatic retries, progress updates, downloads, and saving files into ./assets/chutes/. It also gives connected agents built-in instructions, so they know the workflow when they connect. There is an optional for Claude Code.

It is described as working with Claude Code, Cursor, Cline, Windsurf, Codex, OpenCode, and . Version 1.1.0 has been tested end to end with live image, video, music, and image-editing models, and the project is available on GitHub and npm under the .

Key points

  • It lets generate images, videos, music, and speech inside a project.
  • It reads each model’s live instead of relying on hardcoded settings.
  • It handles sign-in, retries, progress, downloads, and saving generated files.
  • It supports tools including Claude Code, Cursor, Codex, and .
  • It is version 1.1.0, open source, on GitHub and npm, and uses the .
Read original