Which app tells you when to deload, and how it decides
Yes, an app can make this call, and the useful ones make it from your own logged sets rather than from a date in a plan. SayLift does it per exercise: when one lift has gone three or more comparable sessions with no progress and those sessions span 21 days or more, it takes 10 percent off that lift and leaves everything else alone. Before it gets there it has two softer answers, a hold and a 7.5 percent drop, and it tells you which one you are getting and why. This page lists the exact thresholds, what counts as a session worth comparing, and which other apps notice a stall at all.
Updated October 2026
The short answer
Everything below is per exercise. Your bench can go lighter in the same week your squat goes up.
- One slow week is a hold, not a deload. Same weight again, and the coach names the rep count that earns the next increase: last session's lowest working-set reps plus one, capped at the top of your range. In the app it reads "Solid at 60 kg, same weight, push for 10+ reps."
- Reps below your floor means a 7.5 percent cut. If last session dropped under the bottom of your rep range, the next target is 7.5 percent lighter, with the reason "7 reps is below your 8-rep floor, drop back and rebuild." The same cut fires if reps slip across three sessions in a row at the same weight and target.
- A real deload is 10 percent, and it takes three weeks to earn. Three or more comparable sessions with no progress, spanning at least 21 days. The note reads "Three weeks without progress here, take a lighter week and come back fresh."
- It is a per-lift call, not a week off. The Today page says "Go lighter on Bench Press", or "2 lifts go up, 1 goes lighter", with each change shown as previous weight to target weight.
- Coming back after time away is a different case. A 14-day gap on an exercise triggers a restart, not a deload. More on that in getting back to the gym.
- SayLift is $4.17 a month (billed $49.99 a year) after a 3-day trial, and it is iPhone only. For what a deload week actually does to your body, read what is a deload week; this page is about the timing.
SayLift logs your whole workout from one spoken sentence. iPhone only, 873 exercises built in, 3 days free then $4.17 a month, billed $49.99 a year.
Start the free trialWhich apps notice a stall, and what they do about it
Two honest answers exist here. React to the log, which is what SayLift does, or plan the lighter week in advance, which is what Alpha Progression does. The rest of the category hands the decision back to you.
| Notices a stall | What it reads | Per lift or per week | Takes off | Price per year | |
|---|---|---|---|---|---|
| SayLift | Yes, on its own | Your logged working sets for that exercise | Per lift | 7.5% or 10% | $4.17/mo billed yearly ($49.99) after a 3-day trial |
| Alpha Progression | Planned in advance | The scheme it wrote for the block | Per week, scheduled | Set by the plan | $79.99 |
| Fitbod | Not documented | It generates the session, loads included | Not documented | Not documented | $79.99 |
| JEFIT | No | Charts you read yourself | You decide | You decide | $69.99 |
| Hevy | No | Last session's numbers beside the field | You decide | You decide | $23.99 |
| Strong | No | Last session's numbers beside the field | You decide | You decide | $29.99 |
| Liftoff | No | Scores the work after you log it | You decide | You decide | $39.99–$79.99 |
Hold, drop, deload: three answers to a flat week
SayLift's coach returns one action per exercise per session. Three of the six are about things going badly, and the difference between them is how long the trouble has lasted. Knowing which one you are looking at is most of the value, because people reach for a full deload when a hold would have done.
Hold. You are mid-range and stable: you did not reach the top of your rep range, and nothing is falling apart either. The weight stays, and the target becomes a rep count. Lowest working-set reps from last time, plus one, never above the top of your range. This is the first response to a slow week, and plenty of stalls end here, because one more rep at the same load is progress and it resets the clock.
Decrease, 7.5 percent. Two things trigger it. Either last session fell below the bottom of your rep range, which the app names back to you ("7 reps is below your 8-rep floor, drop back and rebuild"), or your reps slid across three consecutive sessions at the same weight with the same target. The cut comes off your last top working weight, so a 60 kg bench becomes 55.5 kg and the point is to rebuild through it rather than sit there.
Deload, 10 percent. This one needs both a count and a calendar. Three or more comparable sessions showing no progress, and those sessions have to span 21 days or more. That second condition is what keeps a bad fortnight from being called a deload. Train a lift twice a week and three flat sessions arrive in under two weeks, which is not enough evidence; the 21-day span makes it a three-week test.
The direction above your rep range is handled on its own page. For how the increases work, and which other apps name a working weight at all, see workout apps that decide your next working weight. If you want the training logic behind reps first and weight second, double progression covers it.
What the coach counts as a comparable session
A plateau is only a plateau if you are comparing like with like, so this part matters more than the percentages. The coach does not simply line up your last few entries for an exercise and look at the top weight.
It compares sessions that had the same rep target and the same number of working sets. Three sets of eight sits next to three sets of eight. Warm-up and drop sets are excluded from that count throughout, so a long ramp-up never makes a session look different from last week's, and a drop set at the end does not count as a fourth working set.
A break of 14 days or more ends the comparison. Sessions on the far side of a gap that long are not evidence about this training block, so the coach stops counting there and treats your return as a restart instead: about 10 percent lighter for one session to find your feet. The dividing line is worth holding on to. Fourteen days away from a lift is a restart. Twenty-one days of training a lift without getting anywhere is a deload. They land on similar weights for completely different reasons.
A session where you actually progressed starts the count again. Add weight or win a rep and the old flat run is over, so a stretch of three good weeks followed by one dull session cannot be read as a three-week stall.
Change your rep target and the coach holds the weight rather than grading you. Move from 5s to 10s and your old numbers say nothing about the new workload, so it keeps the load steady for a session to establish a baseline instead of calling your first set of tens a failure. If you would rather work through a stall by changing the lift or the volume yourself, how to break a lifting plateau walks through the options.
What it cannot see
This is arithmetic on sets you logged. It does not know you slept five hours, that you are eating in a deficit, or that your shoulder has been grumbling since Tuesday. Those are often the real reason a lift went flat, and the coach will read the flat numbers and act on them anyway.
It also reacts rather than plans. The lighter week arrives after three weeks of nothing, which means you spent three weeks finding out. Alpha Progression takes the other route and puts the deload in the calendar before performance drops, inside a block it wrote for you. That is a better fit if you want your training planned, and it costs $79.99 a year against SayLift's $49.99. Our side of that comparison is here.
One practical note on judging your own sets: an app reading rep counts cannot tell a grinding last rep from a clean one. Reps in reserve is the usual way lifters record that, and it stays your judgement to make.
How to set it up in SayLift
- Install SayLift and start the 3-day trial. The coach runs on the phone, so targets appear with no signal.
- Set your rep range if 6 to 12 is not what you train in. The floor is what triggers the 7.5 percent drop, so it should be a number you actually consider a miss.
- Build your routine on the Profile tab, or paste in the plan you already run. The coach follows your routine rather than writing one.
- Log the session, by voice or by hand, and mark warm-ups as warm-ups with the W badge. Warm-up and drop sets are left out of the working-set count, which is what keeps sessions comparable week to week.
- Keep the rep target steady for a few weeks. Changing it resets the comparison and the coach will hold the weight to establish the new baseline first.
- Read the Today page before you start. The headline tells you the shape of the session ("2 lifts go up, 1 goes lighter"), and each exercise shows previous weight to target weight with its one-line reason.
- If a lift gets the 10 percent deload, take it for that lift only and train the rest as normal. One good session back at the lighter weight clears the stall and the next target comes off that.
Who SayLift is for
SayLift suits an iPhone lifter who already has a program and wants the back-off call made from numbers instead of a hunch. You get one instruction per exercise, the threshold it came from, and a lighter week only on the lift that earned it.
It is the wrong app if you want your whole block planned, with the easy week scheduled ahead of time, or if you train on Android. There is no free tier under the trial, no Apple Health sync, and supersets are not grouped, so a pairing goes in as a note.
Pricing is $4.17 a month (billed $49.99 a year), $7.99 a month or $3.99 a week after the 3-day free trial. Get SayLift on the App Store.
Frequently asked questions
How do I know when to deload?
Look at one lift at a time, across several weeks, in a written log. The usual sign is three or more comparable sessions with no extra weight and no extra reps, spread over three weeks or more. A single bad session is not enough, and neither is a fortnight, because two flat sessions in one week can just be a busy week. SayLift applies exactly that test for you and calls it per exercise, so you are not reading charts to work it out.
Is there an app that tells you when to deload?
SayLift does, automatically and per exercise, from the sets you logged. When one lift goes three or more comparable sessions without progress over at least 21 days, the Today page puts that lift 10 percent lighter and gives the reason in a line. Alpha Progression handles it the other way, scheduling the lighter week in advance inside a planned block for $79.99 a year. Hevy, Strong, JEFIT and Liftoff show you the history and leave the decision with you.
How much weight should I take off for a deload?
About 10 percent off your working weight is the common starting point, keeping the same sets and reps. SayLift uses 10 percent for a deload and a smaller 7.5 percent cut for the different case where your reps have just dropped below your rep floor. Both come off the last top working weight for that exercise. The alternatives, such as halving your sets at the same load, are covered in what is a deload week.
Should I deload the whole week or just one lift?
If a single lift has stalled while the others are still moving, there is no reason to make your squats easier because your bench is stuck. SayLift works this way by default, since it decides per exercise, and the Today page will say something like "2 lifts go up, 1 goes lighter." A full week off the gas makes more sense when you feel run down across the board, with achy joints and heavy warm-ups on everything, which is a judgement no app makes for you.
Does an app know the difference between a stall and time off?
SayLift separates them with two different rules. Fourteen days or more without training an exercise triggers a restart, about 10 percent lighter for one session to find your feet again. Twenty-one days of training a lift with nothing to show for it triggers a deload. A gap of two weeks or more also stops the plateau comparison, so a holiday cannot be mistaken for three flat weeks of training.
Do warm-up sets affect the deload decision?
In SayLift, no. Warm-up and drop sets are left out of the working-set count, so the coach compares like with like even if your ramp-up changes week to week, and a long warm-up never drags your target down. They still count toward your volume and tonnage, since the work happened. There is more detail in do warm-up sets count.
Stop typing between sets
Say what you lifted. SayLift writes down every set, rep and weight, then tells you what to lift next time. iPhone only, 3 days free then $4.17 a month, billed $49.99 a year.
Download on the App Store
Reading this on a laptop? Scan to install.