November 6, 2025
Book bot breaks brains
Show HN: I scraped 3B Goodreads reviews to train a better recommendation model
Readers say the robot librarian gets them — now they want it for TV
TLDR: A solo dev trained a book recommender on 3 billion Goodreads reviews, and early users say it nails their taste after a few inputs. The crowd is thrilled, already buying picks, and pushing for a public app connection, open-source plans, and a TV-series version.
A lone dev just dropped a ‘robot librarian’ trained on 3 billion Goodreads reviews, and readers are losing it. Type in a handful of books you’ve read (3+ works best) and it spits out eerily on-point suggestions—only well-known titles show up in results, but it still nails the vibe. One early tester practically squealed, “I’m impressed!” while another immediately clicked buy: Jordan Mechner’s The Making of Prince of Persia is, quote, “on its way to my house.” The site is blazing fast, and with just six books, people are getting picks already on their wish lists. Book nerd joy levels: high.
Of course, the comments didn’t stop at happy squeals. The builders-in-the-chat started chanting API when? open source? and one sharp-eyed reader poked the brain: why choose a “contextual” recommender for books at all—aren’t reading tastes usually context-free? That sparked a friendly geek-off about how much mood and moment matter when you pick your next read. Meanwhile, a crowd is begging for a TV-series version, because binge-watchers want in on the magic, too. Translation: the tool works, it’s fun, and now the internet wants it everywhere. The only real suspense left? Whether the creator shares the toys or keeps the secret sauce locked up.
Key Points
- •Users input books they’ve read to receive recommendations for what to read next.
- •A popularity threshold limits which books appear in search and recommendation outputs.
- •Less popular books are available elsewhere on the site but excluded from recommendations and search.
- •Providing at least three books is recommended for better recommendation quality.
- •The interface requires users to search and select books before generating recommendations.