Movie recommendation tools, by how they actually decide

One of the very few that predicts a score for you specifically, by weighting the users whose ratings have historically correlated with yours. The idea is stronger than the interface, which is showing its age — but the idea is the right one, and it has been right for twenty years.

A GroupLens research project at the University of Minnesota, and the reason a large share of the collaborative-filtering literature exists at all. Rate a few hundred films and the predictions get genuinely good. No streaming integration, no polish — it is a lab, not a product, and it does not pretend otherwise.

Gives you a percentage match per title, computed from your own ratings, for film and TV. Of everything here it is the closest in intent to what I work on, which makes it the fairest comparison and the one I watch most closely.

The social layer for film: diary, lists, reviews, and the best writing community in the category. Following people whose taste you trust beats a global average — but you have to find them first, and what you get is a model of your friends, not of you.

The crowd average at maximum scale, which is both its strength and its ceiling. Unbeatable for settling a fact about a film. But a 7.4 from two million voters still tells you very little about whether you specifically will like it.

Tracking first — it scrobbles what you watch across players and services, automatically. The viewing history it accumulates is excellent raw material for any recommender; the recommendations Trakt itself builds on top are the weaker half.

Tracking across film, TV and anime, with the most forgiving import story if you are migrating from somewhere else. Discovery is secondary to the log.

The open data layer much of this category is built on, including my own project. Thin as a place to browse; indispensable as an API for ratings, credits and localised artwork. Community-maintained, which shows in both the coverage and the occasional gap.

Answers "where can I stream this" rather than "what should I watch". It solves a real problem well and its country coverage is the best here, but it assumes the decision is already made.

Same job as JustWatch — availability and watchlists across streaming services — and strongest on the US catalogue. Tells you where, not what.

Recommendations spanning film, TV, books and podcasts, leaning on social signals and an AI assistant. Broad rather than deep — useful if you want one place for everything, less so if you only care about film.