Video Editor MCP Server — Add Edit Search Videos

Overview

Video Editor MCP Server implements a lightweight Model Context Protocol (MCP) server focused on video content management. It provides a JSON-over-HTTP API to add, update, and search video records and associated metadata, enabling downstream tools (including AI agents and editors) to query and modify the catalog in a consistent way. The server is optimized for workflows where fast lookups and programmatic edits are required, such as AI-assisted editing pipelines, searchable asset libraries, and integration points for media apps.

The server is intended for developers building media tooling or integrations that need a simple, standard interface for video metadata and content references. It is easy to run locally or in container environments and can be integrated into CI/CD, batch processing, or interactive editing apps.

Features

• Add and edit video records (metadata, tags, transcripts, references to stored media)
• Search videos by metadata, tags, transcript text, and custom fields
• Simple JSON API compatible with MCP-style clients
• Lightweight storage configuration (local filesystem or external DB depending on deployment)
• Docker-ready and quick to run locally for development
• Designed to integrate with AI/ML pipelines that need searchable context about videos

Installation / Configuration

Prerequisites: Node.js (16+), Docker (optional)

Clone and install:

git clone https://github.com/burningion/video-editing-mcp.git
cd video-editing-mcp
npm install

Run locally:

# development
npm run dev

# production
npm start

Environment variables (example .env):

PORT=4000
MCP_STORAGE_DIR=./data
MCP_DB_URI=mongodb://localhost:27017/video-mcp   # optional if using MongoDB
JWT_SECRET=replace_with_secure_secret            # if auth is enabled
CORS_ALLOWED_ORIGINS=*

Run with Docker:

docker build -t video-editing-mcp .
docker run -p 4000:4000 \
  -e PORT=4000 \
  -e MCP_STORAGE_DIR=/data \
  -v $(pwd)/data:/data \
  video-editing-mcp

Or using docker-compose (simple example):

version: "3.8"
services:
  mcp:
    image: video-editing-mcp:latest
    build: .
    ports:
      - "4000:4000"
    volumes:
      - ./data:/data
    environment:
      - PORT=4000
      - MCP_STORAGE_DIR=/data

Available Resources

• GitHub repository: https://github.com/burningion/video-editing-mcp
• HTTP JSON API endpoints (see table below)
• Example curl commands and small client snippets (in repo)
• Basic configuration via environment variables and Dockerfile for containerized deployments

API endpoints (summary):

Request and response bodies use JSON. File uploads are usually handled by direct storage references (signed URLs) or multipart endpoints depending on deployment.

Example API Usage

Add a video (metadata only):

curl -X POST http://localhost:4000/videos \
  -H "Content-Type: application/json" \
  -d '{
    "title": "Beach B-roll",
    "description": "Slow motion b-roll of waves",
    "tags": ["b-roll", "beach", "slow-motion"],
    "duration": 12.4,
    "sources": [{"type":"s3","path":"s3://bucket/videos/beach.mp4"}],
    "transcript": "Waves crash softly..."
  }'

Search videos by text and tags:

curl -X POST http://localhost:4000/search \
  -H "Content-Type: application/json" \
  -d '{
    "query": "waves slow motion",
    "tags": ["beach"],
    "limit": 10,
    "offset": 0
  }'

Update metadata:

curl -X PATCH http://localhost:4000/videos/<VIDEO_ID> \
  -H "Content-Type: application/json" \
  -d '{"tags":["b-roll","ocean"], "title":"Ocean B-roll"}'

Use Cases

• AI-assisted editing: An LLM or agent queries the MCP server for candidate clips that match a script or scene description, then returns selected clips for stitching or further processing.
• Searchable media archive: Ingest a large set of clips with transcripts and tags; allow producers to search by phrase, tag, or duration to find appropriate material quickly.
• Content moderation and attribution: Store and update metadata for videos and perform targeted searches to surface content that needs review.
• Automated workflows: CI pipelines that transcode assets and update MCP records when new renditions are ready, enabling downstream apps to react to new versions.

Notes for Developers

• The server exposes a simple, consistent JSON API designed to be embedded into larger systems.
• Consider using an external search index (Elasticsearch, Typesense) for advanced full-text performance at scale; the server can act as the metadata authority while delegating full-text search.
• Secure the endpoints in production (JWT, API keys, or network-level controls) and configure persistent storage (database or object storage) for durability.