Stop working out
what to eat.

It answers “so what do I eat?”

“603 calories left” is not an instruction. Setpoint says 180 g of chicken, 95 g of rice, and the number moves the moment you log something that was not in the plan. Most apps stop at the first half.

An off-plan lunch is not a ruined day

Eat the thing. The remaining meals shrink or grow to absorb it, protein is protected first, and the day still lands on your target. That is the difference between a plan you follow and a plan you abandon by Wednesday.

Australian food, Australian data

Whole foods come from the official Food Standards Australia New Zealand database — the cuts sold here, and entries that say raw, baked, fried or casseroled. Weighing cooked chicken against raw figures undercounts you by nearly a fifth, and most apps never mention which one you are looking at.

Added sugar, not just total sugar

Most apps show one sugar figure, which puts an apple and a can of soft drink in the same number. Eat three pieces of fruit and you read 60 g of sugar and conclude the fruit is the problem — while the guideline you were given was only ever about the added part.

Setpoint shows them separately, and it never guesses which is which. Calling the sugar in fruit and milk intrinsic and everything else added sounds sensible and is wrong in both directions. The added figure comes from the food databases that publish it, and the app tells you how much of your food that covers rather than passing off a partial total as a whole one.

Vitamins and minerals, and how much of them it knows

Thirteen of them — iron, calcium, B12, folate, zinc and the rest — from the same Australian database, for meals out of your plan as well as anything you search or scan. Every figure says what it was built from: from 5 of 7 foods logged. Other apps draw a tidy ring per nutrient, which can only be filled by counting “not recorded” as zero — and a total built that way reads low and reads confident. Ours is less tidy, and does not mislead you.

And when one of them has been genuinely low for a fortnight, Setpoint names it and suggests three foods that would fix it — ranked by how much they carry per calorie, so they fit inside your target, and filtered by what you actually eat. It stays quiet unless enough of your food carried a real figure to be sure.

The AI never invents a number

A model identifies the food and estimates the portion. Every macro figure comes from a nutrition database. If nothing matches, the item is flagged as an estimate rather than quietly counted as fact.

It notices when your expenditure moves

A fall in expenditure caused by being lighter is normal arithmetic. A fall beyond that is worth knowing about, and Setpoint separates the two so you can see which you are looking at rather than assuming the worst.

You can also tell it a stretch of days was not you — gastro, a holiday, a week of antibiotics. Those days are left out entirely rather than fudged, so the estimate says it is less sure rather than pretending nothing was lost.

The scale isn’t telling the whole story

Carbohydrate is stored as glycogen, and glycogen carries water with it. Salt shifts water too. So after a high-carb or salty day the scale can jump even when you have not gained meaningful body fat — gaining a kilogram of fat takes a large sustained energy surplus, roughly 7,700 kcal by the usual estimate, which one evening does not supply.

Setpoint shows you what is likely behind the movement: how much glycogen and associated water your recent carbohydrate could be carrying, alongside the last seven days of your own intake. Instead of seeing “+400 g” and guessing, you can see what changed in the days before it. Scroll back through the diary and it shows the same picture for any previous day.

And once you have logged enough morning weigh-ins, it stops relying on textbook averages and starts measuring your response. Some people see large swings after a high-carb day. Others barely move. Setpoint works out which pattern looks like yours.

The scale is measuring your weight. Setpoint helps you understand it.

Reminders that know what you have already done

A nudge at each meal time you choose, and a weekly weigh-in. A meal you have already logged is never mentioned. No streaks and no daily score — those exist to make you open an app, which is not the same as helping you.

The weigh-in matters more than it looks. It is the measurement your expenditure figure is built from, so a missed one does not just lose a data point, it holds up the number the whole app is for.

One line a day, and only when it is true

Optional, and not a motivational quote. Each morning Setpoint can tell you one thing about your own numbers: how close your expenditure estimate is to being measured rather than guessed, that tomorrow's weigh-in closes out the week it is fitted to, or that you have hit your protein target on twenty-five days now.

On a day when there is nothing true to say, it says nothing. That is the whole design. Every other tracker can send "don't forget to log" — none of them can tell you how many days of your own data the estimate is built on, because none of them measure it.

Trophies for what you did, not what you weigh

Logged thirty days. Twelve weigh-ins. Fifty training sessions. Never a kilo lost — weight is only partly in your control, and an app that rewards eating less is a different and worse thing.

Nothing counts a run of days, so nothing can be broken by a holiday or a week of being ill. And there is no list of the ones you have not got: the wall is empty when you start and fills as you go, because a checklist of things undone is not what a food app should be showing anybody.

Groups, with people you actually know

Start a group, share the code, and compare the month with up to twenty people you invited. Sessions trained, weight moved, days logged, trophies earned — and steps, if you switch them on. First names only, with a chat you can send photos in. Part of the subscription.

There is no public leaderboard and no way to find a group you were not given the code for. Your weight, your body fat, what you ate and how close you came to your calorie target are never shown to anybody, in any group, and there is no setting that would. Ranking people on how little they ate is a competition this app will not run.

Contest prep, inside the app you already use

Dedicated prep apps exist and some of them are good. What Setpoint does differently is run the prep on the same engine as the rest of your year: the expenditure figure it has been measuring from your own logs and weigh-ins for months is the one your prep is planned from, rather than a fresh calculator starting from nothing.

Enter your show date and division and Setpoint works backwards from it. The first thing it does is tell you whether the date is reachable — and if it is not, it shows the arithmetic and the date that is. Finding that out on day one means entering a different show. Finding out at week ten costs a season.

Then one tap writes the whole thing: your targets, a meal plan with however many meals a day you want, and a training programme built for your division. Not three screens and two other tabs — one.

Prep food is expensive and Setpoint says so rather than letting you find out at the till. Your protein target during a prep is the highest the app ever sets, and protein is the costliest way there is to fill a calorie target — so if the week’s shop is going to overrun the grocery budget you set, the prep screen tells you by how much. It does not quietly rewrite the plan cheaper, because the only way to make a prep meaningfully cheaper is to take the protein out. Peak week included — that is the most expensive shop of the whole prep and the worst week to be surprised by it.

From there it is the app you already have, pointed at a deadline: targets set week by week from your measured expenditure, meals and training built against them, and a pace that runs harder early and eases as you get lean — because that is when a deficit starts costing muscle rather than fat.

Cardio is treated as a budget you spend rather than a baseline you set. It starts at an hour a week and climbs — slowly at first, steeply near the end, because the extra work is only worth its recovery cost in the stretch where the deficit can no longer come out of food. The common way to get this wrong is starting high: six long sessions at sixteen weeks out feels productive and leaves you nowhere to go when the scale stalls at week six. Mostly low-intensity, because intervals compete with your lifting for the recovery that is already scarce, and none at all in peak week — that week exists to put glycogen in, and cardio takes it back out.

It is never converted into calories and never subtracted from your food. Setpoint already measures what you burn from what you logged and what the scale did, so the cardio you actually do is in that figure — counting it a second time would cut your food twice for the same work.

Through the middle of it, the question that actually matters is whether you are on the line. Setpoint draws the trajectory your plan expects — a curve rather than a straight line, because losing faster early and easing off as you get lean is the plan, not a deviation from it — and compares it against your weight trend rather than this morning’s reading.

And when you are behind, it shows you what that costs rather than telling you what to do. Two futures, both real: carry on at the rate you have actually been losing and be ready some weeks after this show, or hold the date and here is the rate that takes — including when that rate is past what muscle survives. Setpoint will not pick. It knows the arithmetic and none of the things that actually decide it.

The final week is planned day by day: depletion, the carb load scaled to your lean mass, the training taper, and show day timed backwards from when you are called to the stage. Water loading is explained properly, with the risks at both ends, and you set your own figures — Setpoint tracks them, it does not prescribe them. Diuretics and sweating weight off it will not plan at all.

Then it walks you back up, and it reads you rather than a textbook. Steady weight means room to feed you faster; a real climb eases the pace. It deliberately ignores the first fortnight, because weight after a show rises whatever you do — that is glycogen refilling and bringing its water with it, and treating it as fat would leave you under-fed for weeks on the strength of the one gain the app already knows is not fat.

That stretch decides whether the next twelve months are good or miserable, and it is the part people are most often left to work out alone.

The coach knows about it too, so “why did my calories change?” gets an answer about your show and your week rather than a general one about expenditure.

Fourteen divisions across both sexes, and one for anything a federation has added that Setpoint has not heard of yet — the timeline and the targets work the same either way.

Then one more question: novice, experienced, or masters. A first show gets four extra weeks of runway, because nearly every prep people regret is one that started too late and got made up for with crash dieting in the last month. Masters gets three extra weeks and a lower ceiling on how fast the weight comes off — muscle is harder to hold and harder to rebuild than it was, so the same deficit costs more. If you are over 35 and it is also your first show, pick masters: it is the more careful of the two, and nothing here decides which classes you enter on the day.

Stop it whenever you like. Setpoint works out a fresh target for the bodyweight you are at now — not the one you started on, which is a number computed for a body you no longer have — then rebuilds your meals and your training around it, in the background, while you carry on using the app. A prep programme is not what you go back to, and a prep plan built for a deficit is not either. Everything you have logged stays exactly as it is. And when you finish, it asks the one question worth recording: did you come in flat, about right, or spilled? That answer, the weight you came in at and whatever you wrote down are waiting for you when you set up the next one — and in peak week, which is the only week you can still act on them. A year is long enough to forget what you loaded on. Setpoint does not change anything on the strength of it; it puts last time in front of you and the call stays yours.

Never used if you do not compete. It is one card on the Diet tab, and the rest of the app is unchanged.

Training for a race, not a stage

The same countdown, pointed at a different job. Setpoint works back from the date — but the goal here is arriving able to perform, not arriving at a weight, and almost everything follows from that one difference.

Twenty-eight events across seven disciplines, because not everybody runs. Running from a 5k to an ultra, and trail as its own thing. Walking and trekking — a 5k walk, a half marathon walk, and the 50 and 100 km charity treks that have taken six figures of Australians over a coastline. Cycling, road and mountain and multi-day. Open water swimming, from an ocean swim to a marathon channel crossing. Triathlon from sprint to Ironman, plus duathlon, aquathlon and adventure racing. And fitness racing — Hyrox, DEKA, Spartan, Tough Mudder.

It asks how long you expect to be out there rather than how far you are going, because that is what decides the answers. A four-hour marathoner and a two-and-a-half-hour one are running the same race and solving different nutrition problems — and a 5k walk and a 5k run are the same distance on opposite sides of the line where eating during it starts to matter at all.

It does not rewrite your diet to suit the race. Endurance runs on carbohydrate, and Setpoint will say so if your plan is low-carb and the day is a long one — but saying so is where it stops. Your calories move, because that is what working back from a date means. What you eat and how you split it stays yours, and the app tells you what it would recommend rather than quietly doing it. The race-day fuelling plan applies whatever you eat the rest of the week.

Losing weight is off by default, and the default is the advice. A deficit is a cost paid against the race. Setpoint will run a shallow one early if you ask, and switches it off entirely once the taper starts — glycogen cannot be filled out of a deficit, and the last few weeks are when the training you have already done actually lands.

Then the part most apps have nothing for: what goes in during it. Grams of carbohydrate an hour, scaled to your expected time, with the reason the number changes rather than just the number — above about 60 g an hour one of the gut transporters saturates, so it needs glucose and fructose together. More of a single-source gel does not absorb faster, it just ruins your stomach at 30 km. And the carb load before it, with no depletion phase, because the deplete-then-load protocol is decades out of date.

And then the part that makes the second race better than the first. When you finish, Setpoint asks one question: did you run out, did it hold, or did your stomach turn — and what did you actually take an hour, not what the plan said. Next race that comes back to you instead of the textbook. 60 to 90 grams an hour is a population range, and nobody’s gut is a population. The rate that stayed down in a real race is your own number, and a year later nobody remembers it. Setpoint does not change your plan on the strength of it — it puts last time in front of you and the call stays yours.

It writes the training as well as the meals, and the shape of the week is the point. Most of it is easy — one long session, at most two genuinely hard ones, and everything else at a pace you could hold a conversation through. That is not a soft option, it is the most robust finding in endurance training, and the failure it prevents is the common one: a whole week at a middling effort that feels productive, is too hard to recover from and too easy to adapt to. The plan is built so it cannot write that week.

Strength work stays in rather than being dropped, which is the first thing endurance athletes cut when the hours climb and the thing the evidence is clearest about for staying uninjured. An easier week every fourth, because adaptation happens during the recovery and not the work. And the taper cuts the volume while keeping the frequency — dropping sessions instead feels like resting and loses the feel for the sport in the fortnight it matters most.

Sessions come in minutes and effort rather than sets and reps, and the app logs them that way: a treadmill session asks for time, effort and incline. It will not write a programme for something else — if Setpoint has not been told the sport, a session plan would be invented rather than worked out, and it says so instead of guessing. None of this is coaching. It is a defensible default week, and a coach who knows your history and your injuries beats it comfortably.

The event screen is the prep screen, deliberately. The same countdown in the same place, the day’s calories and carbohydrate under it, the week’s carbs moved towards the days you actually train, what the week’s shop costs against your grocery budget, and one tap each to log your food or open your session. Two different sports, one screen you only have to learn once.

And when the day has been — or when it is not going to happen, which is the more common ending — Setpoint hands you back an ordinary plan. Your calories go back to what the app would set for you now, and the meals and the training are rewritten to match. That matters more here than it sounds: a programme written for a 90 km ride is still a programme written for a 90 km ride the week after you have stopped riding it.

There is no trajectory here and nothing telling you that you are behind. That belongs on a contest prep, where somebody is genuinely racing a body composition. Telling a marathoner six weeks out that their weight is not falling fast enough is not motivation, it is bad advice.

Are you eating enough for what you are doing?

Eating 2,600 calories sounds like plenty. It is not, if two hours of running took 1,100 of them and you carry 55 kg of lean mass — that leaves 27 calories per kilogram to run everything else on, and the line most sports scientists draw is 30. No calorie total on its own shows you that.

That gap has a name and a long list of consequences: stress fractures, cycles that stop, suppressed hormones, and the run of colds that ends a season. Endurance athletes are the textbook case, and almost every tracker will happily watch it happen while showing green numbers all week.

So Setpoint does not use a flat calorie floor for somebody training twelve hours a week. It works out what is left after training, measured against your lean mass, and raises the floor to cover it. It only ever raises one — a watch's calorie estimate is never allowed to lower your target, which is the same rule the rest of the app follows. And if it does not know your body composition it says so and asks, rather than estimating your lean mass and then passing judgement on your fuelling.

A contest prep gets the same number and none of the enforcement, and the difference is deliberate. A prep is the one place in the app where somebody is supposed to spend months under that line — coming in stage-lean is what it costs, and a rule forbidding it would forbid the sport. So the prep screen shows what is left after training and changes nothing about your target. Setpoint does not stop you, does not mark it in red, and does not offer to fix it: the fix is time at maintenance after the show, not a change this week.

It matters most in the last stretch, and that is exactly where nobody is doing the sum. Prep is the one diet where the app itself is pushing cardio up while pulling calories down, month after month — the gap widens on both sides at once. Knowing you are there is worth a great deal more than finding out afterwards, which is how most people find out.

Training that remembers for you

Last session's reps and weights are already in the boxes. Change one digit to beat it. A rest timer sits under every lift, set to that lift's own rest period.

Cardio is not an afterthought bolted onto a lifting app. A treadmill walk asks for minutes, effort and incline — not reps and weight — so a finished session is actually recorded rather than logged as three sets of nothing.

The programme is built to what you ask for, and the choice is an instruction rather than a suggestion. Pick full body, upper / lower, push / pull, push / pull / legs, or one body part a day — and one body part a day means one. Not chest and triceps. Tick the days you can train and it fits the split to your week rather than the other way round, because the days are your life and the split is a preference. Each session lands on its own day — Monday push, Wednesday pull — so the app shows you today’s session, and says so plainly on a rest day. Tick what equipment you actually own, down to “bands, a pull-up bar and a 24 kg kettlebell”, and nothing gets prescribed that you do not have.

Emphasis is its own choice and it is not five settings on one dial. Strength, hypertrophy, compound lifts and conditioning are there, and so are hybrid running and lifting, calisthenics with the actual progressions, joint-friendly work for cranky knees or a return from a layoff, and strength with balance and mobility built through it — the last two being the kinds of training most lifting apps do not reach at all.