looking for an experienced n8n automation developer to build a complete, end-to-end workflow that will:
Workflow Requirements:
-
Ingest long-form published videos from YouTube (via yt-dlp or similar).
-
Detect and extract clips featuring a specific person using AI face recognition (Twelve Labs, Sieve, or similar).
-
Auto-cut highlights based on timestamps, keeping the most engaging moments (positive interactions, key moments, etc.).
-
Reframe to vertical 9:16 format with intelligent cropping to keep the person centered in frame.
-
Add captions automatically (AssemblyAI, Whisper, or similar).
-
Overlay watermark/logo and style captions for social media.
-
Generate thumbnails automatically (Bannerbear, Canva API, or similar).
-
Auto-publish clips to Instagram Reels (Meta Graph API) and optionally TikTok and YouTube Shorts.
-
Log analytics (likes, comments, shares) into Google Sheets.
Tech Stack (Preferred):
n8n (self-hosted on Railway/Docker)
ffmpeg & yt-dlp
AI APIs: Twelve Labs or Sieve (face detection), AssemblyAI or Whisper (captions), Bannerbear or Canva API (thumbnails)
Publishing APIs: Meta Graph API (Instagram), TikTok API, YouTube Data API
Google Sheets API for analytics logging
What I Have Ready:
Starter n8n workflow kit (JSON workflows, Dockerfile with ffmpeg/yt-dlp, .env sample, placeholder watermark)
Hosting environment
Reference image for face detection
What I Need From You:
Connect APIs and credentials
Configure all nodes for smooth end-to-end automation
Ensure error handling and logging
Optional: Set up scheduled posting times in specific time zones
Deliverables:
Fully working n8n workflow chain from YouTube → AI clip detection → edit → publish → analytics
Documentation for setup & maintenance
Short test run with a sample video
Budget: Open to offers based on experience and delivery speed
Location: Remote – global applicants welcome
If interested, please reply with:
Examples of similar n8n or automation projects you’ve built
Your proposed approach for AI detection & clip extraction
Estimated delivery time