AI Book Recommendation Generator

Shellf uses AI text embeddings (not keyword matching) to understand the themes, style, and feel of books in our database. When you select books you've read, we compute a taste profile — a mathematical representation of what you love — and find the books closest to that profile. Then a language model writes personalized reasons for each recommendation.

Yes. You get 2 free recommendation generations per day with no signup required. Each generation gives you personalized recommendations with reasons. Download the Shellf app to get more recommendations and the benefits of library organization and reader insights.

Accuracy improves with more input. With 10 books, recommendations are solid. With 20-30 books, they're remarkably good — the AI has enough signal to understand nuances in your taste like whether you prefer fast-paced thrillers or slow-burn literary fiction. The semantic embeddings catch connections that simple genre matching misses.

Absolutely — this is where embeddings shine. Because the AI understands themes, writing style, and emotional tone (not just genre labels), it frequently recommends books from genres you wouldn't expect. A reader who loves atmospheric fantasy might get a recommendation for literary fiction with magical elements, for example.

Goodreads uses collaborative filtering — 'people who liked X also liked Y.' Shellf uses semantic embeddings that understand the actual content and feel of each book. We also generate a written reason for every recommendation, so you know why a book matches your taste, not just that it does.

Currently, recommendations from this free tool aren't saved between sessions. To save recommendations, track your library, and get even better AI suggestions based on your full reading history and ratings, download the Shellf app (free on Android, iOS coming mid-2026).