Tell me what you’ve seen and I’ll tell you something about yourself.

Why it asks what you have seen, not what you liked

“Have you seen this?” is the cheapest question a film app can ask you. There is no wrong answer, nothing to weigh up, and nobody gets stranded on a film they have never watched — which is exactly where a rating deck loses people. A short deck still carries a real signal, because the films are not chosen at random: some are near-universal and some are the kind you mostly reach by going looking. What you recognise across that split says more about you than a star rating on any one of them would.

Nothing here needs an account. Your taps live in this browser until you decide to keep them, and if you do make an account they come with you, they become the first entries in your diary rather than being thrown away. Tapping a film records that you have seen it and nothing about whether you enjoyed it, because inventing a rating you never gave would sit underneath every average on your taste page afterwards.

Then it guesses, and you mark it wrong

Every recommendation engine asks you to take it on trust, and it can afford to, because you cannot tell whether a suggested film was any good until after you have watched it. A claim about your taste is different, you know within a second whether it is true. So Kinolog makes three guesses about you and hands you the scoring pen: one that is nearly a certainty from what you tapped, one that is a genuine read, and one that is a real reach. You will probably not get three out of three, and that is the point. A “not me” tells it more than a “that’s me” does, and it changes the film you are shown next.

Every claim arrives with its evidence attached, naming the films it came from. A recommender that cannot explain itself cannot be corrected, and one that cannot be corrected never gets better at knowing you. It is also why it says plainly when it is guessing: a first read does not support confident conclusions, and admitting that is more useful than a confident sentence that happens to be wrong.

One film, not a grid

The problem was never that films are hard to find. It is that you scrolled a wall of artwork for forty minutes and went to bed. So the last beat is a single film, with the reason it was chosen, the runtime, and where you can actually watch it, and two alternates underneath in case the first one is wrong. It is weighed against the whole bundled catalogue in a couple of milliseconds, and the page tells you how many films that was and how long it took, because a speed nobody mentions is not a feature.

If you already keep a film diary somewhere

You do not have to start from nothing. A full Letterboxd export imports whole: ratings to the half-star, the date you actually watched each film, your review text, your own tags, rewatches, likes and your watchlist. A decade of history becomes a taste profile in about the time it takes to upload the ZIP. If you are weighing the options, the honest guide to Letterboxd alternatives says plainly where other apps are the better choice.