Video Edit MCP

What video-edit-mcp MCP server does

The video-edit-mcp MCP server provides an MCP interface for common video and audio editing tasks. Video capabilities include trimming, merging, resizing, cropping, rotation, speed changes, fade effects, grayscale, mirroring, text overlays, image overlays, video overlays, transparency, codec-controlled format conversion, frame extraction, and video creation from images.

Audio operations include extracting sound from video, trimming and looping audio, concatenating tracks, changing volume, applying audio fades, mixing multiple tracks, adding audio to a video, and replacing an existing soundtrack. The server also includes video downloads from YouTube and other platforms, directory operations, file listing, and path suggestions for downloads.

How it works

MCP-compatible clients call the server’s editing and utility operations through a standardized interface. Workflows can either produce output paths or keep results in memory. This makes it possible to pass one operation’s result into the next step, such as trimming a source, adding background audio, applying an overlay, and then writing the finished result.

The in-memory workflow is the main reason to use the video-edit-mcp MCP server for multi-step jobs. It avoids requiring a separate intermediate file for every operation and includes utilities for checking stored objects and clearing memory when a workflow is complete. The README’s example uses video information lookup, trimming, audio addition, and a text overlay in one chain.

Setup and configuration

The documented integration requires Python 3.10 or newer. Its listed dependencies include moviepy 1.0.3, yt-dlp 2023.1.6 or newer, mcp 1.12.2 or newer, and typing-extensions 4.0 or newer.

For Claude Desktop or Cursor, the README shows a uvx configuration that selects Python 3.11 and runs the video-edit-mcp package:

{
  "mcpServers": {
    "video_editing": {
      "command": "uvx",
      "args": ["--python", "3.11", "video-edit-mcp"]
    }
  }
}

Claude Desktop stores its configuration in the platform-specific MCP configuration file, while Cursor uses .cursor/mcp.json in a project root. The manual installation path clones the repository, installs requirements.txt, and installs the project in editable mode with pip install -e ..

Tools and capabilities

The video-edit-mcp MCP server covers these supported areas:

• Video editing, effects, overlays, conversion, and frame handling
• Audio extraction, editing, mixing, volume control, and soundtrack replacement
• Downloads from YouTube and other platforms
• Directory and file-listing utilities
• Suggested download paths
• In-memory object storage, inspection, cleanup, and chained processing

The README names Claude Desktop and Cursor as MCP clients, and states that other MCP clients may also use the standardized interface.

Limitations and notes

The provided material documents traditional multimedia operations but lists speech-to-text, text-to-speech, AI audio enhancement, automatic scene detection, face tracking, object recognition, and content analysis as future enhancement areas. Those capabilities should not be treated as currently available features.

The README does not specify hosted execution, authentication, cloud storage, or paid API requirements. Processing requirements and output behavior can vary with the media files and operations selected, but the provided material does not define detailed resource limits or supported codec combinations beyond noting codec control for format conversion.