I built a pipeline that turns every YouTube upload into 8 podcast episodes — with show notes generated automatically. I expected it to take 20 minutes. It took a full day. Here’s what I built, how it works, and the part that genuinely surprised me.
The Problem
Most content creators treat YouTube as their primary channel — but their audience is scattered. Spotify. Apple Podcasts. Amazon Music. Eight platforms, all expecting the same content in different places.
Manually re-uploading, writing show notes, and managing metadata for every episode is hours of work. Per episode. Forever. There had to be a better way.
The Pipeline
Here’s how it works now:
- Upload a video to YouTube as normal
- n8n detects the new upload via YouTube Data API v3
- Modal (serverless compute) strips the audio from the video
- Claude generates show notes and episode metadata automatically
- An RSS feed is updated with the new episode
- All 8 podcast platforms pull from the feed and publish
From upload to distributed across every platform: fully automatic. I touch it once.
The Stack
- YouTube Data API v3 — the trigger that detects new uploads
- n8n — orchestrates every step of the workflow
- Modal — serverless compute for audio processing
- RSS — the universal distribution layer to all podcast platforms
- Claude Code — built the workflow, tested it, and self-corrected it
The Part That Surprised Me
Claude Code didn’t just write the n8n workflow.
It ran the tests. Found the errors. Fixed them. On its own.
I came back to a working, corrected pipeline. The agent had debugged its own output without me stepping in. That’s not just code generation — that’s actual iterative problem solving from an agent.
The Side Build
While I was at it, I also built a thumbnail generator in Google AI Studio. Time taken: 30 seconds.
When the tooling moves this fast, the bottleneck shifts from building to deciding. The question is no longer “can I build this?” — it’s “what should I build next?”
The Honest Take
This wasn’t a 20-minute job. It took several hours. But the pipeline is permanent. Every future episode saves hours of manual work. The maths on automation always wins — you just have to absorb the upfront cost.
If you’re publishing on YouTube and not distributing to podcast platforms, you’re leaving audience on the table. Your listeners are on Spotify, Apple, Amazon, and five other platforms — and you shouldn’t be uploading to each one manually.
Full write-up on Substack: thetop.substack.com
Want to build something like this? Come find us at manyforce.com.
Leave a Reply