Unleash AI-Powered Video Editing with Video Jungle MCP Server 🎬

Effortlessly integrate the power of Large Language Models (LLMs) with the robust video editing capabilities of Video Jungle using the Model Context Protocol (MCP). This server acts as a bridge, enabling you to upload, edit, search, and generate videos through natural language commands.

Experience the magic in action: https://www.youtube.com/watch?v=KG6TMLD8GmA

Prerequisites: To leverage this tool, you'll need an account with Video Jungle and a valid API key.

Core Components

This server provides a comprehensive interface for video manipulation, built around the following key components:

Resources

• Custom vj:// URI Scheme: A unique identifier for accessing individual videos and projects within Video Jungle.
• Project Metadata: Each project includes a descriptive name and detailed description for easy organization and retrieval.
• Intelligent Search: Search results are enriched with metadata, pinpointing specific moments and content within videos, facilitating precise edit generation.

Prompts

(Coming Soon) - Expect future enhancements that will allow for more sophisticated prompt engineering and control over video editing processes.

Tools

The server exposes a suite of powerful tools accessible through natural language:

• add-video: Ingest videos from URLs directly into your Video Jungle library.
  • Returns: A vj:// URI for referencing the newly added video.
• search-videos: Harness the power of embeddings and keywords to search your video collection.
  • Returns: Relevant video matches with detailed metadata, enabling targeted editing.
• generate-edit-from-videos: Create polished video edits from a selection of video files.
• generate-edit-from-single-video: Refine and extract specific segments from a single video source.
• search-local-videos: Search videos from your local Photos app using Apple's tags.

Practical Applications: Real-World Examples

To effectively utilize these tools, ensure you have a Video Jungle account and a valid API key.

1. Adding a Video with add-video

can you download the video at https://www.youtube.com/shorts/RumgYaH5XYw and name it fly traps?

This command downloads the specified video, adds it to your Video Jungle library, and performs a comprehensive multi-modal analysis (audio and visual) for future retrieval.

2. Searching Your Video Library with search-videos

can you search my videos for fly traps?

This query leverages the previously generated metadata to identify relevant video segments, providing a foundation for creating targeted video edits.

3. Searching Local Videos with search-local-videos

Before using this tool, set the environment variable LOAD_PHOTOS_DB=1 to grant Claude access to your local files.

can you search my local video files for Skateboard?

4. Generating a Video Edit with generate-edit-from-videos

can you create an edit of all the times the video says "fly trap"?

This command leverages the context of the current chat to generate a video edit based on the identified segments.

5. Refining a Single Video with generate-edit-from-single-video

can you create an edit of all the times this video says the word "fly trap"?

This tool allows you to extract specific moments from an existing video, creating a concise and focused edit.

Configuration

1. Obtain Your API Key: Log in to your Video Jungle settings and retrieve your API key.

2. Launch Video Jungle MCP:

   $ uv run video-editor-mcp YOURAPIKEY

3. Enable Local Photos App Search (Optional):

   $ LOAD_PHOTOS_DB=1 uv run video-editor-mcp YOURAPIKEY

Quick Start Guide

Installation

Automated Installation via Smithery

npx -y @smithery/cli install video-editor-mcp --client claude

Manual Installation for Claude Desktop

Modify your claude_desktop_config.json file:

• MacOS: ~/Library/Application\ Support/Claude/claude_desktop_config.json
• Windows: %APPDATA%/Claude/claude_desktop_config.json

 "mcpServers": {
   "video-editor-mcp": {
     "command": "uvx",
     "args": [
       "video-editor-mcp",
       "YOURAPIKEY"
     ]
   }
 }

 "mcpServers": {
   "video-editor-mcp": {
     "command": "uv",
     "args": [
       "--directory",
       "/Users/YOURDIRECTORY/video-editor-mcp",
       "run",
       "video-editor-mcp",
       "YOURAPIKEY"
     ]
   }
 }

  "video-jungle-mcp": {
    "command": "uv",
    "args": [
      "--directory",
      "/Users/<PATH_TO>/video-jungle-mcp",
      "run",
      "video-editor-mcp",
      "<YOURAPIKEY>"
    ],
   "env": {
        "LOAD_PHOTOS_DB": "1"
    }
  },

Important: Replace placeholders with your actual directory paths and API key.

Development Workflow

Building and Publishing

1. Synchronize Dependencies:

   uv sync

2. Build Distributions:

   uv build

3. Publish to PyPI:

   uv publish

   • Authentication: Provide PyPI credentials via environment variables or command-line flags.

Debugging Strategies

Debugging MCP servers can be complex due to their stdio-based communication. We highly recommend using the MCP Inspector for an optimal debugging experience.

Launch the MCP Inspector:

npx @modelcontextprotocol/inspector uv run --directory /Users/YOURDIRECTORY/video-editor-mcp video-editor-mcp YOURAPIKEY

Access the Inspector URL in your browser to begin debugging.

Additionally, the server logs detailed information to app.log in the project directory. Utilize logging.info() to add custom logging statements for API call diagnostics.

Monitor the log file in a separate terminal session:

$ tail -n 90 -f app.log