Live
Bookmenti
Every book mentioned in every podcast
Bookmenti finds the books worth reading by listening to the podcasts you already love. It ingests episodes, reads the transcripts, catches every book that gets mentioned, and turns a passing recommendation into something you can browse, search, and jump into at the exact moment it was said. The title that used to vanish thirty seconds after it was mentioned now sticks around.
- What
- Book mentions in podcasts
- Stack
- Bun · Hono · React
- Home
- bookmenti.com
- Status
- Live
The idea that wouldn't die
Bookmenti started years ago as a simple newsletter, and I have rebuilt it more times than I can count. Different stacks, different shapes, each version teaching me something. The core idea never changed: the best book recommendations are buried in podcasts, said once and gone. What changed is that it is finally real, built on a stack (Bun, Hono, Drizzle, React) that actually fits the job.
One database, many storefronts
Here is the part I am most excited about. Books are just the first kind of mention. The same engine that catches book titles can catch anything a podcast talks about: the products, the brands, the people, the stocks, the deals. So Bookmenti is really the opening move in a bigger idea, a whole family of tools fed by one shared transcription database, each one a different lens on the same conversations. Books first, because books are where I started.
Where it is now
Bookmenti lives at bookmenti.com, with over five thousand books caught across a hundred shows and counting. And the engine underneath it has already grown a sibling: Podmenti, where you can read the raw transcripts themselves. It is exactly the kind of project that will probably get rebuilt again as it grows. But that is the fun of it. Some ideas you keep coming back to until they are right, and this is one of mine.
How it works
- Listens to everything
- It pulls podcasts from across the web and reads their transcripts, so nothing said on-air slips past it.
- Understands the mention
- A language model reads each transcript, spots the real book recommendations, and matches them to actual editions in OpenLibrary. Not keyword matching, real understanding.
- Turns talk into a shelf
- Every mention becomes a browsable entry: which book, which show, which episode, the quote, and a link to the moment it was said.
Try Bookmenti
Podcasts are full of book recommendations, said once and gone. Bookmenti listens, catches every mention, and lets you jump to the exact moment it comes up. The first of a whole family of "mentions" tools.