Most fitness apps are diaries. This one makes decisions.
It holds your training and your food in one system — what you lifted, what you ate, what the scale said — and uses all of it to set next week's weights and next week's calories. A diary tells you what you did. This tells you what to do, and why.
There's a free tier. Your log is yours to keep.
FIRST 20 Founding members get the Full tier free for life, updates included — my thanks for getting this off the ground. Take a seat →Next week's weights and rep targets come from what you actually logged this week — not a plan written before it met you.
After eight days of weigh-ins it stops using a formula for your maintenance calories and works out the real number from your own data.
Hate a meal, can't do an exercise, week went sideways — swap it and the day rebalances. A plan you can't live with is a plan you'll quit.
Neither half can answer the question you actually have.
Your body doesn't keep training and food in separate apps, so neither does this. The same weigh-ins and the same food log feed both sides — which is why the app can say things a training app and a diet app could never say to each other.
“Your pressing stalled three weeks while you averaged 280 under target. That's fuel, not programming.”
A food app doesn't know you lift. A lifting app doesn't know what you ate. Neither can produce that sentence, and it's the one that actually changes what you do on Monday. Every claim like it comes with the weeks and the numbers attached — plain arithmetic over your own logs, no black box.
Four days, and four decisions you didn’t have to make.
Tap through the week. The screen is the app; the note beside it is the decision it made and the evidence it made it from.
The weight was already in the field.
You never picked 225. It came from last week: twelve reps at 215, rated just right, top of the band cleared — so the prescription moved and the field was pre-filled before you walked in. 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.
You swapped the chicken. Nothing else moved.
Tap the meal, tap Swap, pick salmon. The incoming food’s grams are scaled so its protein matches what left, and every other item in the meal is byte-identical afterwards — the rice does not quietly move to cover it. Same slot, same calories, same protein. A meal plan you won’t eat is worth nothing, so the app is built to let you refuse any part of it.
Day 8. It stopped estimating your metabolism and measured it.
Eight days of weigh-ins was the threshold. It read your real weight trend against what you actually logged eating over the same stretch and worked backwards: you were averaging 2,410 and drifting down, so you burn more than that. Measured maintenance, 2,640 — not the 2,410 a height-weight-age formula handed you on day one. Every target from here follows your body, and it told you the number rather than quietly changing it.
Three weeks flat on bench while you averaged 280 under target. That's fuel, not programming.
And here is the sentence the whole thing exists for.
It is the first thing on Today, above the session you are about to do. A food app cannot see the bar and a training app cannot see the plate, so neither can produce that sentence at all. This one is a rule over your own numbers with the weeks attached — auditable, defensible to a coach, and not a guess. The intake card underneath it is the evidence: 12% under target, averaged over the days you actually logged this week, and it turns amber on its own once the gap is big enough to matter.
You never have to learn any of this. It works anyway.
Four ideas do the work. If you've been running your own programming you'll recognize every one of them, and you'll want to know they're here — so here they are, named.
Clear the top of your rep range and next week is heavier. Fall short and it holds, or backs off. The decision is made from the session you actually did, every week, without you running the numbers.
After eight days of weigh-ins it compares what you ate against what your weight actually did and works out what you really burn. Then it keeps doing that, because the number moves as you do.
One sentence, in plain words, with the weeks and the numbers attached — and it appears on the lift it's about, when you're about to do it again. Not a wall of charts for you to interpret.
The default asks one easy question after each exercise: easy, just right, or hard. If you train by reps-in-reserve and want that precision, switch it on yourself. Switching it off puts you straight back.
It told me to deload. I tried to train through it. It was right and I was wrong.
This is my own account, exported in September 2026 — six weeks on the cutting program, 446 sets, every number below pulled straight out of it. Going into week 4 the app recommended a deload. I declined it, because I felt fine and the weights weren't heavy. Here is what happened.
Not one plate added. Those four are also the only lifts I rated “hard” all week — the app had the evidence before I'd admit it.
The deload the app wanted at week 4, taken at week 5, after a wasted week of grinding.
1,316 reps across 94 sets — more than any week before it, including the one I refused to rest for.
I'm the one who built it, and I still overruled it. It didn't argue and it didn't quietly override me — it made the call, logged that I declined, and let me find out.
That's the whole design in one week: it decides, it tells you why, and you're still the one holding the pen. A spreadsheet would never have raised it. A coach might have, at the next check-in, a week after it mattered.
Plenty of apps say they're built on the research. Fewer will tell you which parts aren't.
Where the literature gives a real answer, the app uses it and the working is written down: weight-loss rates are held under the ceiling Helms and colleagues set out for natural lifters, deload depth follows the tapering meta-analyses, protein targets follow the consensus range.
Where it doesn't, the internal specs say so in those words — one progression rule is recorded as “a design preference with a coherent rationale, not a research finding.” Judgment calls are labeled as judgment calls.
That's the part I'd rather be held to. Anyone can cite a study. Telling you where the evidence runs out is the harder promise, and it's the one that should decide whether you trust the number in front of you.
A coach can suggest. Only you can accept.
If you train with a coach, they see your real numbers instead of screenshots you remembered to send. What they can't do is quietly change anything. Every prescription arrives as something you approve, and the rules live in the database, not in a promise.
It runs both ways. While a coach is overseeing your program you can't silently restructure it either — a mid-week swap goes to them as a request. Nobody finds out after the fact that the plan changed.
If you part ways, one tap ends the link — and you keep every single thing you ever logged.
Access is limited to a coaching link you accepted, and it ends when the link is revoked or expires.
Not your logs, not your history. Changes to your program come to you as a request.
No one outside an accepted coaching relationship can see your data. We don't sell it, and we don't use it for advertising.
The honest competition isn’t another app.
It does exactly what you tell it, which is the problem: every progression decision is still yours, every week, forever. And it doesn't know what you ate.
Here: the arithmetic runs itself, and shows its work.
It adds up what you ate and stops. The target it gave you on day one came from a formula and usually never moves again.
Here: the target is measured from your own trend.
A plan somebody wrote before they met you, that doesn't change based on what happened when you did it. Big libraries, deep content — genuinely their strength.
Here: fewer templates, but the plan recalculates weekly.
The coaching is the good part. The doc is the bad part — screenshots, stale numbers, and a plan that ages between check-ins.
Here: same coach, real data, no screenshots.
Every one of them is a snapshot. A plan from January is a guess by March.
Each one holds a decision made on a day that has passed. Then you get stronger, your maintenance calories move, a bad month happens — and the numbers sit there unchanged until someone rewrites them. Usually that someone is you, on a Sunday night. Here it happens on its own, every week, which is why this is a subscription and not a plan you buy once.
The lock only appears when there's something real behind it.
Every app you've used dangles a locked feature at you whether or not it has anything to show. So you learned to ignore them.
Here a locked card only renders when the engine actually found something in your data. A quiet week shows no lock at all. When one does appear it carries the real evidence — “a pattern links your food log and your pressing stall, three weeks of data support it” — and only the reading is behind the gate.
Which makes the lock itself worth something: it never bluffs, so its appearance means your data contains something worth reading.
Recording the past is free. Prescribing the future is paid.
Every account logs food, weigh-ins and workouts with the full tooling — barcode scanner, real serving sizes, portion fractions. Nothing is crippled to push an upgrade. What the paid tiers add is interpretation. And if you log free for six months then upgrade, the intelligence has half a year of your history to work from on day one.
Log everything — every set, every meal, every weigh-in. Full fidelity, no caps. It's your training diary and it's yours to keep.
Now the app coaches your lifting. It picks your next weight and rep target from what you actually logged last week.
Everything above, plus the food — meals built from what you actually eat, and calorie targets measured from your own body instead of a formula.
Full tier, plus me. I build your program, I see your logs between sessions, and the app flags you before you have to tell me.
Get it on Google Play, or open it in your browser.
On an Android phone, install it from Google Play like any other app. On an iPhone or a computer there's nothing to download: you open a link, sign in, and add it to your home screen. After that it opens from an icon and runs full screen, exactly like any other app on your phone.
It has to be Safari — Apple doesn't allow other iPhone browsers to install apps this way. Takes about ten seconds and you'll never do it again.
Search for ITLabs Fitness, or use the Google Play button at the top of this page. Same account and same data as the website.
Handy for planning a block or doing the week's shop. It's the same account and the same data as the phone.
I'd rather have twenty people I can talk to than a thousand I can't.
I built this because I was running my own training out of a spreadsheet and my clients' out of documents, and neither one could tell me the thing I most wanted to know: whether the problem this month was the program or the food.
Everything on this page is built and running against real training. What it has not met is a room full of strangers, and strangers break software in ways people you know never will — so the founding group stays small on purpose. Twenty people I can actually talk to while they use it, who’d rather be early and heard than late and one of thousands. They get the Full tier free for life for taking that on. If that’s you, put your name down and tell me how you train.
When it breaks, I'd like to hear it from you before I hear it from a review.
The first twenty get it free, for good.
My thanks for being the people who got this off the ground.
Access opens November 4. Ask now and you're in the first group through the door.
Your address is used to contact you about the beta. Nothing else.
This probably isn't for you if…
Twenty seats, and I'd rather they went to the right people. Genuinely no hard feelings if one of these is you — I'd just be the wrong tool.
There are excellent apps for that and they're cheaper. Half of what you'd be paying for here is the training side talking to the food side.
That's a real thing to want, and it isn't this. The whole argument here is that next week's plan should be different because you are.
You'd be fine here eventually — but the first twenty are people I'm learning from, not people I'm teaching. Come back when the onboarding has been through a few rounds and you'll have a much better time.
Things will break, and when they do I'll want to hear about it from you rather than read it in a review. If that sounds like a chore instead of the point, this isn't the right month.
The ones people actually ask.
Do I have to train to failure?
No, and the default mode never asks you to. Hit the top of your rep range and the weight goes up; rate how the exercise felt and that shapes your next block. There's an opt-in RIR mode for people who specifically want a proximity-to-failure ramp — you turn it on yourself, and turning it off puts you straight back.
Is it AI? Can I trust what it tells me?
No black box. Every insight is plain arithmetic over your own numbers, and it shows the evidence — which weeks, which figures. If the app can't defend a sentence with your data, it doesn't say it.
What happens to my data if I cancel?
The paid features switch off. Everything you logged stays readable forever, and you can export it yourself from inside the app without asking anyone. Features leave, data stays — that's a hard rule, not a policy we might revisit.
I don't lift. I just want to lose some weight.
Then you'll mostly feel the nutrition side: targets that adjust to your real metabolism, meals you can swap without wrecking the math, and a weekly note on how your trend is actually going. The lifting depth is there when you want it, and quiet until then.
My schedule is a mess. Does that break it?
No. Sessions aren't pinned to calendar days — the next one is whatever you haven't done yet, and progression comes from your last logged session rather than from the date. Move a day, miss a day, train at 4am; the math doesn't mind.
Why a subscription and not a one-off plan?
Because you're not paying for the plan. You're paying for next week's plan being different because the body it's for is different — recalibrated from your own logs rather than reissued unchanged. Buy a plan once and you own a snapshot of a person who no longer exists.