Hey n8n community ![]()
I wanted to share how you can use n8nβs built-in MCP Client Tool to connect to the Ssemble AI Clipping MCP server and automate short-form video creation from YouTube videos β no custom nodes needed.
What is Ssemble AI Clipping?
Ssemble is a video processing platform that takes YouTube videos and automatically creates TikTok/Reels/Shorts-style clips with:
- AI-selected highlight moments with viral scoring
- 9+ caption templates (Karaoke, Bold, Minimal, etc.)
- 45+ background music tracks
- 30+ gameplay overlays (Minecraft, GTA, Subway Surfer, etc.)
- 159+ meme hook clips (2β5 second attention grabbers)
- AI-generated titles, descriptions, and hashtags
We recently published an MCP server that exposes all of this as 9 tools any MCP client can use.
How to connect it in n8n
Since n8n has native MCP support via the MCP Client Tool node, setup is straightforward:
Step 1 β Prerequisites
- Self-hosted n8n (v1.88.0+)
- Set environment variable:
N8N_COMMUNITY_PACKAGES_ALLOW_TOOL_USAGE=true - Get a Ssemble API key from: https://app.ssemble.com (Settings β API Keys)
Step 2 β Add MCP Client Tool node
-
Add MCP Client Tool node
-
Transport: SSE
-
URL:
https://mcp.ssemble.com/mcp -
Authentication: Header Auth
- Header Name:
X-Ssemble-API-Key - Header Value:
sk_ssemble_your_key_here
- Header Name:
-
Tools: All (or select specific)
Step 3 β Connect to AI Agent
- Add an AI Agent node (Claude, OpenAI, etc.)
- Connect MCP Client Tool as a tool input
- The agent now has access to all Ssemble tools
Available tools (9 total)
| Tool | Description |
|---|---|
create_short |
Submit YouTube URL for clipping (returns request ID) |
get_status |
Check progress (0β100%) |
get_shorts |
Get clips + download URLs + viral scores |
list_requests |
List all requests |
delete_request |
Delete a request |
list_templates |
Browse caption styles |
list_music |
Browse music tracks |
list_game_videos |
Browse gameplay overlays |
list_meme_hooks |
Browse meme clips |
Example workflows
1. Content repurposing pipeline
RSS β detect new YouTube video β create_short β wait/poll β get_shorts β post to Slack/Discord
2. Batch processing
Google Sheet (URLs) β loop β create_short β collect request IDs β poll β save results back
3. AI-powered automation
Webhook β user input (βmake it funny with gaming overlayβ) β AI Agent chooses tools β generate clips β return results
Important: Async processing pattern
Video processing takes 5β30 minutes, so:
create_shortβ returnsrequestIdinstantlyget_statusβ check progressget_shortsβ fetch results
Tip: Add a Wait node (5β10 min) or loop with polling until completed.
Links
- npm: https://www.npmjs.com/package/@ssemble/mcp-server
- GitHub: GitHub - ssembleinc/ssemble-mcp-server Β· GitHub
- MCP Endpoint: https://mcp.ssemble.com/mcp
- n8n Docs: MCP Client Tool node documentation | n8n Docs
Would love to see what you build with this ![]()
Happy to answer any questions!