Full courses + unlimited support: All my FREE resources: 14 day FREE n8n trial: Code NATEHERK to Self-Host n8n for 10% off (annual plan): In this video, I walk you through a no-code RAG (Retrieval-Augmented Generation) workflow I built using n8n. I’ll show you how I take YouTube video transcripts, store them in a Supabase vector database, and enrich them with metadata like video titles, URLs, and timestamps. This allows the RAG agent to tell me exactly where the retrieved answer came from — including which video, the link to that video, and even the moment in the video it pulled the data from. I also break down what metadata is, why it matters, and how it dramatically improves retrieval accuracy and context. If you’re building any kind of AI assistant or agent, this will be a foundational concept to understand — and I make it as simple and practical as possible. Sponsorship Inquiries: 📧 sponsorships@ WATCH NEXT: TIMESTAMPS 00:00 Quick Demo 01:18 Why Metadata Matters 03:35 RAG Pipeline w/ Metadata 08:59 Download This FREE Workflow 10:02 Metadata Filtering 11:54 Automatic Deletion Pipeline 13:55 Final Thoughts 14:37 Want to Master n8n? Gear I Used: Camera: Razer Kiyo Pro Microphone: Blue Yeti USB