Comfy AE Video Factory MCP

An MCP Server for AI short video automation pipelines: connecting ComfyUI image-to-image/image-to-video, FFmpeg post-processing, After Effects template packaging, review, and multi-platform publishing preparation.

Design Goal: Enable Cursor / Claude Desktop / ChatGPT MCP clients to call these tools to automatically complete the "Generate → Package → Review → Export → Publishing Package" workflow.

Capabilities

• Invoke local or cloud ComfyUI workflows

• Query ComfyUI task history

• Analyze videos using FFprobe

• Transcode, crop, extract frames, and concatenate using FFmpeg

• Generate JSX scripts for automatic After Effects asset replacement

• Invoke aerender to render After Effects templates

• Create standardized video production task manifests

• Generate publishing package configurations for Douyin / Xiaohongshu / Bilibili / YouTube Shorts / TikTok

• Provide basic content review placeholder interfaces for future integration with third-party review APIs

Related MCP server: levea-mcp-server

Installation

npm install
npm run build

Running

npm start

Configuration in MCP Clients

{
  "mcpServers": {
    "comfy-ae-video-factory": {
      "command": "node",
      "args": ["/absolute/path/comfy-ae-video-factory-mcp/dist/index.js"],
      "env": {
        "COMFY_LOCAL_URL": "http://127.0.0.1:8188",
        "VIDEO_FACTORY_ROOT": "/absolute/path/workspace",
        "FFMPEG_BIN": "ffmpeg",
        "FFPROBE_BIN": "ffprobe",
        "AERENDER_BIN": "/Applications/Adobe After Effects 2025/aerender"
      }
    }
  }
}

Windows Example:

{
  "AERENDER_BIN": "C:\\Program Files\\Adobe\\Adobe After Effects 2025\\Support Files\\aerender.exe"
}

Recommended Workflow

1. pipeline_create_job to create production task directory and manifest

2. comfy_submit_workflow to submit ComfyUI image-to-video workflow

3. comfy_get_history to query results

4. media_probe to check video specifications

5. ffmpeg_transcode to unify aspect ratio, bitrate, and frame rate

6. ae_generate_jsx to generate AE template replacement scripts

7. ae_render_template to invoke aerender and output the final packaged video

8. review_video_package for review placeholder

9. platform_prepare_package to generate publishing packages for various platforms

AE Template Integration

It is recommended to pre-create these placeholder layers in your AE project:

• {{MAIN_VIDEO}}

• {{TITLE}}

• {{SUBTITLE}}

• {{LOGO}}

• {{CTA}}

Then use the JSX generated by this project to replace assets and text, and output via aerender.

Security Notes

This MCP Server executes local commands, so the following restrictions are applied by default:

• All input and output files must be within the VIDEO_FACTORY_ROOT directory

• FFmpeg parameters are encapsulated via a whitelist rather than directly exposing arbitrary shell commands

• AE rendering supports dryRun; it is recommended to test commands first

GitHub Upload

git init
git add .
git commit -m "init comfy ae video factory mcp"
git branch -M main
git remote add origin git@github.com:YOUR_NAME/comfy-ae-video-factory-mcp.git
git push -u origin main

Remarks

The official MCP TypeScript SDK supports creating servers, registering tools, and connecting to MCP clients via stdio or Streamable HTTP. This implementation uses stdio, which is suitable for local automation integration.