Methodology
How we rate books
Every book card on RecomendaLibros shows a consensus rating across multiple public sources. We don't invent numbers — and we'd rather show you fewer reviews than fake ones.
Sources we draw from
For each recommendation our literary AI cites the public-domain rating data it knows about, primarily from these platforms:
- Goodreads. The largest reader-driven rating site (≈90M users). Our model summarises Goodreads' average score and review count for each title.
- Amazon. Customer reviews aggregated across Amazon's worldwide marketplaces. Heavily weighted because purchase-verified reviews are harder to game.
- OpenLibrary / crítica especializada. Public catalog of editions and a small but growing community rating layer. Acts as a sanity check.
How the consensus is calculated
1. The AI estimates each source's average rating and approximate review count for the book. 2. We display the AI-cited consensus average (the score under the 5 stars) and the SUM of approximate review counts as the total signal. 3. We never scrape Goodreads or Amazon — both forbid programmatic access in their Terms of Service. Numbers are based on the model's training data and may be a few months stale. 4. Books that score below 3.8 average are filtered out before they reach you.
An honest disclaimer
Counts shown are best-effort estimates from the AI's knowledge cutoff. They are NOT live numbers. If you want the freshest data, click any book and read recent reviews on Amazon or Goodreads directly. We're working on integrating live OpenLibrary aggregates next.
Last updated: April 2026 · v1.0