I listen to the Lex Fridman podcast a lot, but finding a specific topic across episodes is not very convenient.
You can search YouTube, of course, but that is not quite the same thing as searching through podcast chapters in a clean and focused way. Sometimes I do not want recommendations, comments, or unrelated results. I just want to find where a topic shows up and jump to that part.
So I built lex-fridman-podcast-search.
The idea is simple: collect podcast data, make it searchable, and let the interface focus on chapters rather than the usual YouTube browsing experience. It is basically a distraction-free way to search through the Lex Fridman archive and move around episodes more quickly.
The project also ended up being a nice excuse to put together a “trend” backend stack at the time with NestJS, Apollo GraphQL, Prisma, MySQL, and Redis. There is also a live version at lexpodcast.vercel.app.
This is probably a niche project, but I like niche projects when they solve a real annoyance. In this case, the annoyance was simple: long podcast episodes are useful, but they are much more useful when you can search them by topic and chapter instead of hunting manually.