Chessy

The diagnostic corner of this directory is crowded - Aimchess, ChessDNA, RV Chess and Chess.com's own Insights all read a player's games and report back on what is going wrong. They share a shape: a dashboard of charts and scores that the player has to interpret and then act on. Chessy's bet is that the dashboard was never the hard part. It opens on a chat window. The player types "review my last game" or "what should I work on", and the answer arrives as prose with an interactive board or a statistics card embedded in the reply. Tap a number on the card and the games behind it open. There is no menu to learn.

That is a real difference, and it is worth being precise about what it is not. The marketing attaches "No other chess app does this" to the Sunday coaching report - on the website and on both store listings - and the claim is wrong: Aimchess has shipped weekly study plans for years, and the puzzles-built-from-your-own-blunders sitting next to it are ChessDNA's territory as much as anyone's. The claim is also unnecessary. Strip the superlative and what remains is still the most frictionless version of this idea on a phone: the analysis layer is ordinary, the interface on top of it is not.

The cost of putting a language model in front of the analysis is the obvious one, and it is the question this directory put to the publisher directly. The answer came back more substantial than most of the category offers: every position is analysed with Stockfish, an in-house layer builds the teaching and the reasoning out of that engine output, and the finished explanation is checked against Stockfish once more before the player sees it - the model is not evaluating the board by itself. That is the right architecture. It is also, as of September 2026, entirely unpublished. Nothing on the site, in either store listing, or inside the app tells a player any of it, and there is no audit and no visible indicator separating an explanation that passed its check from one that only reads as though it did. Chess Pangolin, listed here the same month, built its coach out of deterministic board checks and publishes a scored honesty audit - not because its engineering is necessarily better, but because a claim a reader can verify is worth more than one they have to take on trust. The 1.5.8 release notes are unusually candid about where the gap led ("the coach can no longer say 'here they are' over an empty space"), which says the failure mode was real and is being worked on.

The other caveats are age and price. Chessy arrived on iOS in April 2026 and Android in May, carries 15 App Store ratings and 100+ Play installs, and is written by one person who answers the support email. Everything visible suggests that person is working hard - shipping monthly, writing honest release notes, answering an editorial query with specifics instead of a slogan. None of that makes $79.99 lifetime a safe bet five months in. And unlike every other app in this corner, Chessy keeps nothing outside the paywall: the download is free, the trial runs three days on iOS and seven on the Play yearly plan, and then the app stops, so that trial is the whole evaluation window. The annual remains the sane way to test whether talking to a coach is a habit that survives the novelty. For a player between roughly 600 and 1800 who already plays online, already knows the blunders are the problem, and has never once opened the analytics tab they were given, the question is not whether the diagnosis is better than Aimchess's, but whether being able to ask for it out loud is what finally makes it get used.