Ssemble AI Clipping MCP server

Ssemble MCP Server

MCP server for Ssemble AI Clipping — create viral short-form videos from any AI assistant.

Works with Claude Desktop, Claude Code, Cursor, VS Code, Windsurf, and any MCP-compatible client.

What it does

This MCP server lets AI assistants create short-form video clips from YouTube videos using Ssemble's AI clipping engine. The AI can:

• Create shorts from YouTube URLs or uploaded files
• Webhook callbacks — get notified when processing completes (works with n8n, Zapier, Make)
• Browse assets — caption templates, background music, gameplay overlays, meme hooks
• Track processing — check progress and retrieve completed clips
• Manage requests — list history, get results, delete old requests

Prerequisites

• Ssemble account with an active subscription
• API key (get it from Settings → API Keys in the Ssemble dashboard)
• Node.js >= 18

Quick Start

Claude Desktop

Edit your config file:

• macOS: ~/Library/Application Support/Claude/claudedesktopconfig.json
• Windows: %APPDATA%\Claude\claudedesktopconfig.json
• Linux: ~/.config/Claude/claudedesktopconfig.json

{
  "mcpServers": {
    "ssemble": {
      "command": "npx",
      "args": ["@ssemble/mcp-server"],
      "env": {
        "SSEMBLE_API_KEY": "sk_ssemble_your_key_here"
      }
    }
  }
}

Restart Claude Desktop after saving.

Claude Code

claude mcp add ssemble -- npx @ssemble/mcp-server

Then set the environment variable:

export SSEMBLE_API_KEY="sk_ssemble_your_key_here"

Cursor / VS Code / Windsurf

Add to your MCP settings:

{
  "mcpServers": {
    "ssemble": {
      "command": "npx",
      "args": ["@ssemble/mcp-server"],
      "env": {
        "SSEMBLE_API_KEY": "sk_ssemble_your_key_here"
      }
    }
  }
}

Available Tools

How Processing Works

Video processing takes 5-30 minutes. Two workflows available:

Option A: Polling (manual check)

1. create_short — submits the job and returns instantly with a request ID
2. get_status — check progress anytime (queued → processing → completed)
3. get_shorts — retrieve the generated clips when done

Option B: Webhook (automated)

1. create_short with webhookUrl — submits the job with a callback URL
2. When processing completes or fails, the API sends an HTTP POST to your URL with the results
3. No polling needed — ideal for n8n, Zapier, Make, and custom backend integrations

Webhook payload example:

{
  "event": "shorts.completed",
  "requestId": "507f1f77bcf86cd799439011",
  "status": "completed",
  "timestamp": "2026-03-30T12:00:00.000Z",
  "data": {
    "shorts": [
      {
        "id": "...",
        "title": "AI-Generated Title",
        "video_url": "https://...",
        "duration": 45,
        "viral_score": 8.5
      }
    ]
  }
}

Usage Examples

Create shorts from a YouTube video:

"Create shorts from https://youtube.com/watch?v=abc123, use the first 10 minutes with chill background music"

Browse available assets:

"Show me all caption templates" / "What background music is available?"

Check progress:

"What's the status of my video request?"

Shortened. The full README is on GitHub.