Building a learning library with Claude + Notion — podcasts, articles & videos
Remember all those podcast episodes and articles you saved for later? I do.
They're in browser tabs, Spotify queues, YouTube's Watch Later, and screenshots of recommendations I took at 11pm. The intention is always there. The system, not so much. This time, I built a learning library in Notion, with Claude doing the logging so there's minimal friction from logging a useful or interesting piece..
The problem isn't lack of discipline — it's that logging things takes just enough effort to break your flow. Even a quick task (copy a title, open Notion, fill in the fields, pick a tag) is a context switch. And the more steps between you and getting back to what you were doing, the less likely you log anything at all.
The strategy is to not rely on sheer will.
I wanted something where I could hand off the work to Claude — paste a link or a title, say "log this," and be back to whatever I was doing in seconds. Not a recommendation engine. Not a content library someone else curates. Just a place for everything I'm actively reading and listening to, easy to add to, and easy to come back to.
There's a name for what happens without a system like this — the collector's fallacy: saving and gathering feels productive, but without a way to revisit and process, it's just hoarding with good intentions.
I started with podcasts — a batch of episodes I'd been meaning to go through. I shared them with Claude and asked it to log each one into my Notion reading list, with title, author, content type, tags, and a status field. End to end, it took about two minutes. The manual version — opening each link, copying titles, pasting into Notion, setting properties — would have taken 15 to 30 minutes.
Claude also noticed the tag I wanted didn't exist in the Notion database yet, so it created the tag in the schema and applied it to all entries in one pass.
[INLINE IMAGE — to add: A real screenshot of the Notion reading list database showing several logged podcast entries, with visible columns for title, type, tags, and status. Should look like a working tool in use — not a product demo. Crop to show 4–6 rows with real data.]
Where it breaks — and the workarounds
Not everything went smoothly. Claude's Spotify connector can't resolve individual podcast episode URLs — it doesn't pull episode metadata from a link the way you'd expect. The workaround: on Spotify's desktop app, grab the embed code (the <iframe> snippet) and paste that to Claude instead. Claude fetches the public embed page and extracts the episode title from there. No Spotify connector needed — it's just reading a URL.
The catch: this only works on desktop. Spotify's mobile app doesn't expose embed codes, so logging from your phone still means typing the title manually.
For articles and web pages, the path is simpler — Claude can fetch a URL directly and pull what it needs. I haven't tested YouTube or Apple Podcasts links yet, but those are on the list.
The podcast test was proof of concept for the workflow: find something worth learning from → hand it to Claude → it shows up in Notion, tagged and ready. The real value comes when this covers everything I learn from, not just one format.
Next sources I want to test: articles and blog posts (URL to Notion entry with summary and tags), YouTube videos (can Claude pull title and channel from a link?), Apple Podcasts links (does the same embed workaround apply?), books (title or ISBN to Notion entry with metadata), and Vimeo videos (less common, but worth checking if the URL fetch works).
The system is simple enough that each of these is a small experiment, not a rebuild. The library grows format by format, and Claude handles the repetitive parts each time. That's kind of the whole point — small effort to log, easy to come back to, and it actually sticks.