What using it actually looks like.

You pick a program, answer a short setup, and start training. The weights it puts in front of you on day one are estimates from its exercise database — they are meant to be wrong, and you are meant to correct them. Change a weight mid-set and it asks whether that’s for today or for the rest of the program.

Weigh in every morning even when it feels pointless. Eight days is the threshold where the app stops estimating your calorie needs from a formula and starts measuring them from your own weight trend against what you actually ate. The earlier you cross it, the sooner every target on the food side is yours rather than a population average.

Open the session and it runs one lift at a time. The lift name is the biggest thing on the screen, because at arm’s length mid-set that’s the only thing you need to be sure of. Under it: what to do today, and what you did last time you were here.

The weight field is already filled in from your last session. The reps field is deliberately empty — a prescribed rep count sitting in a box becomes a logged rep count on one tap, and then your history claims lifts you never made. Type what you got, tick the set, and the rest timer starts on its own.

When every set of an exercise is confirmed, it asks one question: easy, just right, or hard. That answer plus what you actually hit is what decides next week. It writes the recommendation as a sentence you can read — it never applies one behind your back, and it never moves you to the next lift without you saying so.

Your meals are planned, not typed. Each one is built from foods you said you’d actually eat, solved in grams so the day lands on its macro targets. Most days the only thing you do is tick the meals you ate.

If you ate some of it, tap the meal and say how much — three quarters, half, a quarter. If you skipped it, mark it skipped. A skipped meal is the one state that dims itself, because it means the thing genuinely didn’t happen, and the day’s totals should say so.

Tap the meal, tap swap on the thing you won’t eat, and pick something else. The replacement is scaled in grams so it lands where the old one did on the macro that mattered, and nothing else in the meal moves.

This is the least discoverable thing in the app and the most important. A meal plan you won’t eat is worth nothing, so refusing part of it is a supported move, not a failure.

Nothing expires and nothing is pinned to a calendar date. The next session is whatever you haven’t done yet, and when you open one the app starts you at the first station with unfinished work.

Progression reads your last logged session, not the date on it. Train at 5am on a Saturday, skip a Tuesday entirely, come back after ten days away — the arithmetic doesn’t mind, and you never start a week already behind.

Swap it. The replacement takes the same slot in the day, so the rest of the block is untouched and your next session still knows where you are. If the new lift has no history, the app says so in plain words rather than showing you a blank — a first week on a movement is a normal state.

Every few weeks the app will look at your logged work and recommend backing off. It will say why. You can decline it — it records that you did and carries on.

It is worth taking seriously. The evidence it uses is your own dropping rep counts at unchanged weights, which is the thing you are least able to see from inside a hard week.

Everything you log is yours and stays readable whether or not you’re paying. You can export it yourself from inside the app without asking anyone, and you can delete your account yourself too.

If you stop subscribing, the parts that interpret your data switch off. The record doesn’t. Features leave, data stays.