In this showcase, I demonstrate how you can build a RAG setup specifically for searching over Youtube content and a n8n-hosted frontend to use it.
Hey there 
I’m Jim and if you’ve enjoyed this article then please consider giving me a “like” and follow on Linkedin or X/Twitter. For more similar topics, check out my other AI posts in the forum.
I’ve been really enjoying the n8n @ scale - Office Hours LiveStreams (thanks Angel & Marcus!) lately and if you haven’t seen it, I highly recommend checking out them out if you’re considering a more serious move to n8n for your team or organisation.
The livestream format is highly interactive with questions collected prior to the show but also a lot more pouring in from the live chat. I am noticing however that there were quite a few repeat questions which were answered in the previous stream. This is not necessary the audience’s fault as there will always be a flux of new people and we can’t expect them to know which episodes to catch up on.
Typically, you would advise the viewers with repeat questions to search until you realise Youtube channel search is actually quite bad! Frustrated, I decided to try build my own Youtube search on the livestream transcripts using a simple RAG approach in n8n.
Initially, this template only searched the n8n @ scale office hour livestream recordings but I later decided to extend it to other videos across the channel. This turned out to be a great way to explore n8n videos I haven’t watched yet!
Check out the demo here: https://jimleuk.app.n8n.cloud/webhook/n8n_videos
The Stack
Apify for Youtube Transcripts (Apify.com)
As you may know, I’m a big fan of Apify for their range and quality of their scrapers. The Youtube scrapers are particularly good - really easy to use and can help overcome some of the official youtube API usage limitations.
(Psst! My promo code is still active until the end of May 2025! Use 20JIMLEUK to get 20% off any subscription for 2 months)
Qdrant for Vector Store (Qdrant.tech)
Still my favourite vector store due to the awesome search and filtering APIs that come built-in. For this template, I used Qdrant’s Search Groups API to great effect - this ensures more videos get shown in the results compared to naive similarity search.
N8N for everything else! (n8n.io)
One cool feature of n8n is how it is also able to handle both simple backend and frontend tasks. Alongside the vectorise workflow, this template also includes a 2 webhook triggers; 1 to serve the Web UI to perform searches and 1 which acts as an API endpoint to return search results dynamically.
The Template
I’m releasing this n8n template for free and you can grab a copy on my creator page. Let me know if you run into any issues!
Requires n8n version 1.80+
Conclusion
Building a Youtube RAG workflow isn’t new I’d admit but a powerful skill to add to your AI arsenal. Hopefully this template can shed some light on how you can build your own using n8n!
Also big thanks to the n8n team for the putting together the quality videos and streams for the community 
Keep up the great work!
if you’ve enjoyed this article then please consider giving me a “like” and follow on Linkedin or X/Twitter. For more similar topics, check out my other AI posts in the forum.
Still not signed up to n8n cloud? Support me by using my n8n affliate link.
Need more AI templates? Check out my Creator Hub for more free n8n x AI templates - you can import these directly into your instance!