$9.99 to $19.99/month · 3-day free trial · AI workout planner and gym logger · Progressive resistance training, dumbbells, barbells, kettlebells or machines . Personally tested: 3 sessions across the 3-day trial on iPhone · knee issue declared at onboarding · exercise swaps, added exercises, Planning tab and BodyScan · Prices verified September 2026
FitnessAI is an AI strength-training app that prescribes the sets, reps and load for every exercise and adapts them as you log. It suits an experienced lifter who wants the weekly decisions handled. The injury question it asks at signup does not change what it puts in front of you. Prices start at $9.99 a month with a 3-day trial, and I tested it over three sessions with a knee problem declared.
FitnessAI Review 2026: Quick Answer
Quick Verdict
Worth it for women over 40? Yes, if you already lift and want the weekly programming handled for you. FitnessAI scores 5.5/10: a strong progressive strength engine, without the midlife extras some rivals add.
Strength training is the part of a midlife routine that earns its place, and it is the part FitnessAI gets right. It scores 7.5/10 for Muscle Potential because it prescribes a load for every set and raises it as you log, so progressive overload happens whether or not you remember what you lifted last week. Time Efficiency scores 8/10, its joint highest mark with UX and Design. I opened the app, followed what was on the screen and finished, with nothing to decide and no plan to write.
What it does not do is anything beyond the lifting. There are no injury modifications, no perimenopause or menopause programming, nothing that asks how you slept, which is why Recovery Compatibility scores 3.5/10, and effectively nothing for stress or mind-body work. None of that is unusual in an AI strength app. It lands in the same band as Fitbod at 5.8/10 and below Caliber at 7.3/10, where the extra money buys you a human coach.
Recommended for an experienced lifter who owns dumbbells and a bench, wants her training week decided for her, and is confident enough to swap a movement she does not like. Not the right fit if you need to be taught the movements, if you are training around a joint problem, or if you want the menopause side of midlife training addressed.

Download from the App Store and subscribe there, so Apple handles the billing.
- Price
- $9.99 to $19.99/mo, $59.99 or $89.99/yr
- Free trial
- 3 days, auto-renewing
- Note
- Not every plan includes a trial. Check your own paywall.
Is FitnessAI right for you?
FitnessAI is strongest for building bone density (7.0 out of 10, against a 5.1 average across the 87 programmes scored) and weakest for training on low energy (1.5 against 5.7).
| Question | Verdict | Rating | Score |
|---|---|---|---|
| How it suits a perimenopausal body | |||
| Is FitnessAI good for pelvic floor problems?Based on pelvic floor support and low-impact design | Low, and worse than mostmost score 4.7/10 | 2.0 / 10 | |
| Does FitnessAI include perimenopause and menopause education and programming?Based on programming, education, hormone awareness and credibility | Low, and worse than mostmost score 3.2/10 | 0.5 / 10 | |
| Is FitnessAI good for painful joints?Based on low-impact design and recovery provision | Low, and worse than mostmost score 6.2/10 | 3.0 / 10 | |
| Is FitnessAI manageable when your energy is low?Based on recovery provision and mind-body support | Low, and worse than mostmost score 5.7/10 | 1.5 / 10 | |
| Is FitnessAI good if you are new to strength training?Based on beginner on-ramp and mind-body support | Low, and worse than mostmost score 6.1/10 | 2.5 / 10 | |
| Is FitnessAI good for building bone density?Based on bone loading and progressive strength | Strong, and better than mostmost score 5.1/10 | 7.0 / 10 | |
| Does FitnessAI help with nutrition?Based on midlife nutrition support | Low, and worse than mostmost score 3.8/10 | 0.0 / 10 | |
| How it performs as a product | |||
| Can you fit FitnessAI into a busy week?Based on time efficiency | Strong, about the same as mostmost score 7.5/10 | 8.0 / 10 | |
| Is FitnessAI worth the money?Based on value for money | Low, and worse than mostmost score 6.9/10 | 4.5 / 10 | |
Jump to Section
FitnessAI Review 2026: Why I Tested It
I have tested somewhere between forty and fifty fitness platforms over the last fifteen years, and I have trained through perimenopause myself, so I am not the beginner this kind of app usually worries about. I am not a clinician, and nothing here is medical advice. Where a claim needs clinical backing I have cited the research and named it in the text.
I wanted to know whether an app that decides your sets, reps and load for you is worth paying for in midlife. Progressive strength work is the highest-value thing most women over 40 can add, and the usual reason it stops is not motivation. It is the admin of deciding what to lift each week.
So I tested the part FitnessAI sells. Does it take those decisions off you, does it progress you sensibly, and what does it leave you to sort out yourself? I trained three sessions on the trial, with a knee problem declared at signup, and logged everything it put in front of me.
What Is FitnessAI?
FitnessAI is an AI workout planner and gym logger for resistance training, made by FitnessAI Inc. and now owned by the app studio Appex Group, aimed at anyone who lifts weights. It generates a weekly strength plan from a signup questionnaire, prescribes the exact sets, reps and load for every exercise, and adjusts those numbers based on what you log.
It is not a class library and there are no guided videos. Each exercise appears as an animated anatomical illustration with the muscles you are working shaded in blue, above a written Pro Tip and a numbered list of steps. You read the form cues rather than being talked through them. There is no coach, human or virtual, in the loop at any point.
Which FitnessAI is this?
Who is behind FitnessAI?
FitnessAI was founded in 2019 by Jake Mor, a software engineer, and went through the Y Combinator accelerator in the winter 2020 batch. Y Combinator now lists the company’s status as acquired.[3] It is owned and operated by Appex Group, a Boston studio that buys and scales mobile apps and lists more than fifty in its portfolio.
I could find no coach, physiotherapist, physician or exercise scientist named anywhere: not on the FitnessAI website, not on either store listing, not in the help centre, and not on the parent company’s site. The blog authors carry job titles but no qualifications. This is a data and engineering product and it is open about this, with a store listing claiming the algorithm “is sure to out-perform any human personal trainer or fitness coach at the gym”.[4]
That claim sits awkwardly beside the company’s own guidance elsewhere, which is that you should override the algorithm when it is wrong: “If you notice a weight that you know will be too light or too heavy for you, change it!”[5]
How I Tested FitnessAI
I tested FitnessAI over three sessions across the 3-day free trial, on an iPhone, training at home with dumbbells and kettlebells, with a knee problem declared at onboarding. The app put me on a three-day rotation and I did all three: a chest and triceps day, a back and biceps day, and a legs and shoulders day. I used the exercise swaps, I added extra exercises, I went through the Planning tab properly and I ran a BodyScan. I did not test it over enough weeks to judge how the progression behaves across a training block, and I say so wherever that matters below.
The onboarding questionnaire
Signup is a chat-style questionnaire of roughly fourteen steps, presented as messages from a “Fitness AI coach”. It asks what you are hoping to achieve, what keeps you motivated, how you would describe your fitness level, your age, your gender, which areas you want to focus on, whether there is anything it should know about physically, what equipment you have and how often you want to train.
The injury question is the one that matters here. It offers post-covid recovery, “can’t do the jumps”, back or hernia, arms and shoulders, hip joints, knee, or none of the above. The equipment question offers full gym, barbells, dumbbells, kettlebells, machines or none. I picked Knee, and dumbbells plus kettlebells.
What it never asks about: menopause status, perimenopause symptoms, HRT, pelvic floor, or anything to do with your cycle. It takes your age and your gender and then does nothing with either beyond the numbers. I checked the website, both store listings, the help centre and the company blog afterwards, and the words perimenopause and menopause do not appear anywhere.
One thing did make me raise an eyebrow. Under “What keeps you motivated to exercise?”, alongside health and wellness, weight management and stress relief, sits “Become sexually attractive”. It tells me something about who the product pictures when it pictures its user, and I do not think it is a woman in her fifties trying to hold onto her bone density.
The first workout, and the weight it chose for me
FitnessAI prescribes a specific load for every exercise, and its first guess at that load can be a long way out. My back and biceps session came out as a dumbbell row at 11kg (about 24 lb) for 8 reps across 4 sets, a high shrug at 6kg (13 lb), a curl at 7kg (15 lb), a reverse curl at 3kg (7 lb) and a side bend at bodyweight. Some of those were sensible. Others were nowhere near. It opened one lift at 2kg (4 lb) when I train with 8kg (18 lb) in each hand, and it prescribed a backwards lunge at 0kg, which is to say bodyweight.
To its credit, the app is upfront about this. Every unfamiliar exercise carries a Learning Phase notice reading: “This is the first time we are recommending this exercise to you. Please adjust the weight and reps as needed.” So it is not pretending to know. It hands the calibration back to you and learns from what you log.
What I did not see, in three days, is anything that tells you when to go up over the following weeks. You are told to correct today’s number. You are not coached on progression. That is a real limitation for anyone who does not already have a sense of what a hard set feels like, and I come back to what the evidence says about light opening loads in the science section below.
There is also a feature I have not seen elsewhere and rather liked. Tap into an exercise and you get Standings, which benchmarks you against other users filtered to your age band and your bodyweight. On the incline dumbbell press it showed me 186 users, aged 40 to 50, weighing 61 to 68kg (135 to 150 lb), averaging 17kg (37 lb) for 8 reps. That is a far more useful reference point than a generic average.
Less to my taste is Champions, a per-exercise leaderboard with named users and their lifts. Mine sat at the bottom with a shrug emoji next to it. Some people will find that motivating. Others will find it exactly the thing they were trying to get away from.
Swapping and adding exercises
Swapping exercises is the feature I would keep if I could keep only one. Every exercise opens into an Alternatives row: equipment-matched swaps you can drop in with one tap, each shown with the same animated illustration and the same muscle shading.
I learned new movements from it. Seeing four ways to load the same muscle group, side by side, with the working muscles shaded, taught me more in three sessions than a lot of longer programmes have. You can also add exercises on top of what it gives you, which is how my sessions grew from the twenty to thirty minutes I usually do to about forty.
The catch is in that screenshot. Look at the third alternative offered for a backwards lunge: a plyo squat, a jumping movement, offered to a user who declared a knee problem at signup on a screen that also offered “Can’t do the jumps” as a separate option. I deal with that properly in the women over 40 section below, because it is the finding this review turns on.
The Planning tab, which is where it gets good
The Planning tab is where FitnessAI stops being a workout generator and becomes something you can shape. I nearly missed it, and it would have been a shame.
At the top you build the week: Day 1 chest, triceps and abs, Day 2 back, biceps and abs, Day 3 legs and shoulders, add a fourth day if you want one, and assign muscle groups per day. Underneath sit the controls: routine, duration, objective, equipment, rep ranges, weekly goal, experience level, rest timer, a randomness slider, quick logging, circuit training, pause mode, gym detection and a reset for the training algorithm.
That rest timer is worth flagging, because it is often described elsewhere as fixed. Mine was set to a 20 second to 1 minute range and was adjustable. The randomness slider, sitting at 50%, is a stranger inclusion: it is how the app decides how much to vary your exercise selection, and I suspect it is behind the frequent complaint that the programming feels arbitrary.
BodyScan
BodyScan estimates your body composition from your phone camera, using pose estimation. It walks you through the capture with a video demonstration and a skeleton overlay, and the company’s launch announcement pitches it as a way to scan at home “for a fraction of the price of a DEXA Scan”.[6]
I could use it inside my trial without being asked for more money. I am not going to tell you it is included, because FitnessAI publishes no price for it anywhere and I only saw it on a trial, and other users report being asked to pay extra for the feature on a paid plan. Check what your own paywall says before you count on it.
I want to be careful about the number it gives you, too. A camera-based estimate is not a DEXA scan, and I would not make training decisions on it. Some users report readings well out of line with a trainer’s measurement. Treat it as a rough marker to watch over time rather than a measurement.
Challenges
The Challenges tab runs time-boxed group programmes with prize pools. During my trial the recently completed ones were “Dad Bods in Motion”, a 30-day challenge with 425 participants, and “Muscle Mami Season”, 21 days with 515. Prizes included a $250 Nike gift card and a lifetime FitnessAI subscription. There was no live challenge running while I was in the app.
These are the one place video appears: the challenge intros are a person speaking to camera, though the workouts inside them are the same animated illustrations as everywhere else. The points system is worth reading closely. A completed workout earns 10 points, a daily streak 5 and a weekly streak 50, but inviting a friend earns 20 and posting about FitnessAI on social media with a tag earns another 20 a day. Two of the five ways to climb the leaderboard have nothing to do with training.
What Equipment Do You Need?
A pair of dumbbells and a bench are the realistic minimum, and adjustable dumbbells are the upgrade that matters most with this particular app. FitnessAI will generate a plan for a full gym, but the version most women reading this will use is the home dumbbell version, which is what I tested. Here is what I would own before starting, based on the loads the app prescribed me.
It also has an Apple Watch app and syncs to Apple Health. I logged from my iPhone throughout and did not test the Watch app across a training block, so I cannot tell you how well it holds up as the primary logging device.
Starter dumbbells (up to 10 lb)
Recommended. Curls, raises and the accessory work sat here.
Medium dumbbells (up to 25 lb)
Recommended. The workhorse pair. My loads ran 3kg to 11kg in one session.
Adjustable bench
Recommended. Pressing and supported rows come up in the rotation.
Resistance bands
Recommended. The app programmes no warm-up, so bring your own.
Adjustable kettlebell
Optional. Kettlebells are selectable at signup if you have one.
A quick heads-up: some of the buy links above are Amazon affiliate links, so if you grab something through them I may earn a small commission, at no extra cost to you. I only ever point you to kit I’d happily use myself.
Is FitnessAI Good for Women Over 40?
As a strength tool, yes, if you already know how to train. As a menopause-aware product, no, and like most strength apps it does not claim to be. The distinction matters, because training appropriate for a midlife body and menopause-specific programming are two different things, and FitnessAI has some of the first and none of the second.
What it gets right for this age group
If I had to pick one thing worth doing in midlife, I would pick progressive resistance training, and FitnessAI is built entirely around it. It prescribes load, it tracks what you lift, it pushes the numbers up. A meta-analysis of twelve randomised trials in healthy postmenopausal women found large improvements in both lower and upper body strength from resistance training, and supported a frequency of three days a week in 60-minute sessions.[7] Three days a week is exactly what the app put me on.
Standings is the one feature here I would call age-aware, and it is age-aware by accident rather than by design: the cohort filter happens to give a midlife lifter a realistic reference point instead of a chart built on twenty-five-year-olds. I have described what it showed me in the first-workout section above.
Where it has limits
There is no recovery input of any kind. You cannot tell it you slept badly, that your joints are aching this week, or that you are in a stretch where everything feels harder than it should. Recovery is the thing I have found changes most through perimenopause, and it is the variable the app has no concept of. Users training over longer blocks report the same thing, and I have set out what they say in the reviews section below.
There is no deload, no rest-day programming, no warm-up and no cool-down. There is no pelvic floor work and no acknowledgement anywhere that pelvic floor exists. And there is nobody with a clinical or coaching qualification associated with the product to have thought about any of it.
It does not modify for injuries, and it says so. I declared a knee problem at onboarding and was still offered a dumbbell plyo squat as an alternative to a backwards lunge. That is documented behaviour rather than a fault: the help centre answers “Can the AI take my injury into consideration?” with “The AI doesn’t automatically account for an injury given injuries can be really complex and varied. However, you can set it up to meet your needs while injured.”[1] The company blog points you to the Alternatives row to make the swap yourself.[5]
No comparable app does better. I checked Fitbod, Ladder, Alpha Progression, Caliber and Dr. Muscle in September 2026 and found no automated injury handling in any of them. Caliber and Ladder can route it through a human coach, the rest leave it to you. The practical note is that the Alternatives row tells you what a swap trains but not what it demands, so if you are working around a joint, treat each suggestion as a proposal to check. Exercise is recommended for knee osteoarthritis by every current international guideline, though the most recent Cochrane review judged the short-term benefits to be of uncertain clinical importance.[8]
This is the perimenopause and menopause assessment, and it is separate from the overall score. Twelve things a programme either does for a midlife body or does not, each scored out of 10 with my reasoning next to it. Together they produce a single menopause fit score, and that becomes one of the nine categories in the overall scoring below.
How FitnessAI Fits Perimenopause, Menopause & Beyond
Women Over 40 is our most important criterion, and here's the full breakdown behind it.
FitnessAI is strongest for progressive strength, bone loading and low-impact, and weakest for menopause programming, pelvic floor, hormone awareness, menopause education, midlife nutrition, mind-body support, recovery & deload, expert credibility, low-impact and beginner on-ramp, in our menopause and midlife capability assessment for women over 40.
At a glance
What the app teaches about perimenopause and/or menopause, not whether the training suits a midlife body.
| Capability | Score | Rating | Why |
|---|---|---|---|
| Right training for a midlife body | |||
| Low-impact | 4/10 | Minimal | Dumbbell strength by default, but plyometric swaps appear in the alternatives list even after a knee is declared. |
| Progressive strength | 7/10 | Strong | Core function: prescribes load, sets and reps and adapts on logged performance. |
| Bone loading | 7/10 | Strong | Programmes progressively loaded compound lifts, which is the mechanism that builds bone. No impact work, and the opening loads run light until you correct them. |
| Recovery & deload | 2/10 | None | No recovery input, no deload, no rest-day programming, no warm-up or cool-down. |
| Menopause education and programming* | |||
| Menopause programming | 0/10 | None | None. No menopause or perimenopause programming anywhere in the product. |
| Pelvic floor | 0/10 | None | None. Pelvic floor is never mentioned. |
| Hormone awareness | 0/10 | None | None. Age and gender are collected at signup and never used for this. |
| Menopause education | 0/10 | None | None. No menopause education in the app, site, help centre or blog. |
| Expert credibility | 2/10 | None | No coach, physiotherapist, physician or exercise scientist named anywhere. Founder is a software engineer; owner is an app-acquisition studio. Point awarded for the real 40k-lifter dataset. |
| Whole-woman support | |||
| Mind-body support | 1/10 | None | Competitive leaderboards and a 'become sexually attractive' motivation option; no mind-body content. |
| Beginner on-ramp | 4/10 | Minimal | Every exercise carries a written Pro Tip and a numbered step-by-step description, which is more than most apps give you. Held back by there being no video and by the app asking you to correct its own load suggestions. |
| Midlife nutrition | 0/10 | None | None. An AI Calorie Tracker cross-promotion for a sibling app is not nutrition integration. |
ScaleBest-in-classStrongSomeMinimalNone
What the sources say: Injury policy, cancellation routes, trial terms and lifetime tier verified against FitnessAI's own help centre; pricing pulled live from the US App Store; ownership confirmed via Y Combinator and Appex Group. Caveats: 3-day sample test only, exercise library size not published by the company, BodyScan pricing not published.
* What the app teaches and programmes around menopause, not whether the training suits a midlife body. Many programmes that score low here are still a good fit, they simply never use the word. A high score is not proof either: a few menopause videos added to a general library lifts this score without changing the training.
These are evidence-informed suitability ratings, not clinical measurements. They combine exercise science with product-specific research and, where indicated, hands-on testing, alongside multi-source verification, and are approximate. Individual response varies. A starting point, not medical advice. How we score menopause fit
FitnessAI Weighted Scoring: How the 5.5/10 Was Calculated
This is the overall score, and it asks a broader question: how good is this programme, full stop. Nine categories, each weighted by how much it matters to a woman in perimenopause or beyond, which is why Muscle Potential counts for more than Value for Money. The Women Over 40 Specificity row is the menopause fit score from the twelve capabilities above, carried through. Scores are based on three sessions inside the 3-day trial plus documentary research, so this is a sample-tested review rather than a long-run one.
| Category | Weight | Score | Weighted |
|---|---|---|---|
| Time Efficiency | 15% | 8.0 | 1.20 |
| Muscle Potential | 18% | 7.5 | 1.35 |
| Women Over 40 Specificity | 10% | 2.5 | 0.25 |
| Joint Friendliness | 12% | 3.5 | 0.42 |
| Recovery Compatibility | 10% | 3.5 | 0.35 |
| Programme Structure | 12% | 6.5 | 0.78 |
| Value for Money | 8% | 4.5 | 0.36 |
| UX and Design | 8% | 8.0 | 0.64 |
| Nutrition Integration | 7% | 1.5 | 0.11 |
| Total | 100% | 5.5 / 10 |
Scoring notes: Time Efficiency 8/10 and UX and Design 8/10 are the high marks, and they are earned. The app is fast, the animations teach, the alternatives carousel is excellent and the Planning tab is deeper than it first looks. Muscle Potential 7.5/10 recognises a real progressive overload engine, held back by opening loads that need correcting, an exercise library the company will not size, and users reporting erratic progression over longer blocks. Women Over 40 Specificity 2.5/10 is the composite of the twelve capability scores above. Joint Friendliness 3.5/10 reflects an injury question that does not filter anything, by the company’s own account. Recovery Compatibility 3.5/10 and Nutrition Integration 1.5/10 are the floor, because neither exists in the product.
Final Weighted Score
5.5 / 10
Good at removing decisions, poor at accommodating a midlife body, and it asks about your joints without using the answer
The Science: Does a Starting Weight That Is Too Light Matter?
Less than you would assume for building muscle, more than you would assume for building strength, and it depends entirely on how hard you take the set. This is worth understanding before you dismiss FitnessAI over the 2kg suggestion, because the answer is more interesting than “light weights are useless”.
A meta-analysis of 21 studies found that gains in one-rep-max strength were significantly greater with heavy loads than light ones, but that changes in muscle size were similar across a wide spectrum of loads.[9] A later network meta-analysis of 28 studies in 747 adults reached the same conclusion: hypertrophy appears to be load-independent, while strength gains are superior with heavy loads.[10]
Here is the condition that matters, and it is the one FitnessAI does not manage. Both of those analyses included only studies where sets were taken to or close to failure. A light set stopped at a prescribed rep count well short of failure is the combination the evidence does not cover. FitnessAI gives you a rep target and no concept of how hard that set should feel, no reps in reserve, no rating of exertion. So if you accept a load that is too easy and stop at the prescribed number, you are doing the version the evidence does not cover.
Progression also does not have to mean adding weight. An eight-week trial comparing adding load against adding repetitions found both to be viable strategies with differences of questionable practical significance.[11] So the practical answer for a FitnessAI user is simple: when the suggested weight is too light, either put the weight up, as the app explicitly invites you to, or keep going until the set is genuinely hard.
Two more findings are worth having in mind for this age group. Lean mass declines specifically around the menopause transition, not merely as a function of getting older.[12] That is the reason I think the timing of starting to lift matters. And the LIFTMOR trial showed that heavy resistance and impact training improved bone density in postmenopausal women with low bone mass, with only one adverse event across the study.[13] That trial is also the clearest illustration of what FitnessAI is missing: participants were screened for conditions affecting bone and function, and every session was supervised. An app that states it does not account for injuries cannot replicate either condition.
Finally, in the app’s favour: a randomised trial comparing app-guided resistance training against therapist-supervised training in older adults with sarcopenia found no significant differences in outcomes between the groups, though a 58-person, four-week trial is too small to demonstrate equivalence.[14] The delivery mechanism is not the weakness. An app is a perfectly reasonable way to deliver strength training. The question is what the app knows about you.
What Do Other Users Report?
The headline rating and the current mood are two different things, and the gap is worth knowing about before you subscribe. FitnessAI carries a 4.7 star average on the US App Store, built from roughly 56,000 ratings accumulated since 2018. When I read the 147 most recent US App Store reviews, close to half of them were one star. Recent reviewers always skew towards the motivated and the aggrieved, so do not read that as the truth about the app. Read it as a sign that something changed after the ratings were banked.
The complaints that come up again and again
Billing and cancellation dominate everything else. About thirty of those reviews mention cancelling and roughly a dozen mention refunds. Reading them closely, three separate problems are being described. The first is the one this review can settle: people who cannot find the subscription in their iPhone settings, which is what happens when you buy through the website rather than the App Store. One reviewer put it plainly: “It isn’t handled by Apple so can’t cancel in the normal way through phone settings.” The second is being charged before the trial ended, often with the promised 24-hour reminder never arriving. The third is BodyScan: several reviewers say they bought a plan they understood to include it and were then asked to pay again. I cover how to get out of both billing channels in the cancellation section below.
The paywall on basic logging. “There’s no way to log exercises or do anything unless you pay”, and one person describing being stopped mid-workout: the app “wouldn’t let me log the set”.
Repetition and a limited exercise library. A twelve-year lifter wrote that the app “feeds the same exercises” and called it “very cookie cutter”, and another said they have to “edit and replace some exercises every single day”.
Settings not being respected, which corroborates my knee finding. One Google Play reviewer, writing in November 2024, said the app “totally ignored the info input on set-up with fitness level and body issues”. Others report equipment selections resetting themselves and bodyweight exercises reappearing after being turned off.
Progression logic going the wrong way. Several people report the algorithm recording successful sessions as failures and reducing their weights, and one describing assisted pull-ups and dips where the app increases the assistance weight after a success, which makes the exercise easier rather than harder.
Adverts inside a paid app. One reviewer wrote “PLEASE STOP SHOWING ME ADS IN AN APP I PAY FOR”. That is what I would expect from a product owned by a studio with more than fifty apps to cross-promote, though I saw no adverts myself inside the trial.
Android is a worse experience than iOS, with repeated reports of BodyScan not working at all and, at one point, saving a workout being broken.
What the positive reviews say
They cluster on exactly the thing I liked. It removes the decisions, it is simple, and the progression feels automatic. One reviewer wrote that they had had multiple personal trainers over the years and never seen results like they had with FitnessAI. Another, giving five stars, wrote the line I keep coming back to. Describing the weight and rep increases, she said they are “too linear and don’t take age into consideration”. That is a fan of the app describing precisely the limitation I scored it down for.
User sentiment above is drawn from App Store and Google Play reviews read in September 2026. It reflects what people chose to write, not verified fact, and I have not independently confirmed any individual account.
FitnessAI specifications
- Free trial
- 3 days on the plan shown, card required, auto-renews. Not offered on every planSource: https://fitnessai.zendesk.com/hc/en-us/articles/22926865660059-Why-was-I-charged-immediately-free-trial
- Cancellation
- Easy if bought in-app via Apple (Settings > Subscriptions > FitnessAI > Cancel). Website-billed subscriptions do not appear in iPhone settings and must be cancelled in-app under Membership. Fees non-refundable except where required by lawSource: https://fitnessai.zendesk.com/hc/en-us/articles/23319971089179-Website-subscription-vs-Apple-Store-subscription
- Session length
- 20 to 90 min setting; app recommends 40 to 60; 40 min in testingSource: In-app Duration control, September 2026
- Who makes it
- Founded 2019 by Jake Mor, a software engineer (Y Combinator W20). Acquired; now owned and operated by Appex Group. No coach, physiotherapist, physician or exercise scientist named on the site, either store listing, the help centre or the parent company siteSource: https://www.ycombinator.com/companies/fitnessai
- Still maintained
- Active; iOS v23.2 (2026-09-07), Android updated 2026-09-01
- What is included
- On-demand classes: noLive classes: noRecovery content: noMenopause education: noStructured programmes: partlyMeditation: noNutrition guidance: no
- Women over 40
- Pelvic floor work: noMenopause clinician involved: noInjury modifications: noBone-health focus: noModified progressions: noCycle syncing: noPrenatal options: no
- Devices and access
- Offline downloads: noCasting to TV: noSubtitles: noWearable sync: yes
- Support
- Clinical advisory board: noCommunity: partlySupplement upsell: no
- Last verified
- 7 September 2026
How Much Does FitnessAI Cost?
There is no single price, and that is the answer. On the day I checked, the US App Store listed ten separate live subscription products for FitnessAI, including three different monthly prices and two different quarterly prices, all running at the same time. Whatever you are shown at signup is very likely not what the person next to you is being shown.
| Option | Price (US) | Notes |
|---|---|---|
| Weekly | $3.99 | The most expensive way to pay by a wide margin if you stay more than a month. |
| Monthly | $9.99, $14.99 or $19.99 | Four separate monthly products are live simultaneously, at three different prices. A 2x spread for the same app. |
| Quarterly | $24.99 or $44.99 | Two quarterly products, also live at the same time. |
| Annual | $59.99 or $89.99 | Three annual products, at two different prices. I was shown two different annual prices in my own account, matching this pattern. |
| Lifetime | Not sold | FitnessAI’s help centre states plainly that it does not offer a lifetime membership, and no lifetime product appears in the App Store purchase list. A lifetime subscription has been given away as a challenge prize, which may be why some users believe one is for sale.[17] |
| BodyScan | Not published | FitnessAI publishes no price for BodyScan anywhere. It was available to me inside the trial without a further charge, but I cannot tell you whether that holds on a paid plan, and users have reported being asked to pay extra for it. Confirm at signup. |
| Free trial | 3 days on my plan | Auto-renewing. The company’s own help centre notes that some plans do not include a trial at all.[2] |
Apple displays a maximum of ten in-app purchases, so this list is not necessarily exhaustive, and the UK store shows a different ladder again including an “Early Bird” product with no US equivalent. The practical advice: read the price on your own paywall carefully, and do not trust a figure quoted in any review, including this one, as the price you will be offered.
There is effectively no usable free tier. Multiple users report being unable to log a set without subscribing, which puts FitnessAI behind Alpha Progression and Caliber’s free tier, both of which give away substantial functionality for nothing.
How Do You Cancel FitnessAI?
If you subscribed through the App Store, it takes about ten seconds. If you subscribed through the FitnessAI website, it is a different and much worse experience, and that distinction explains almost every complaint you will read.
I subscribed through Apple and cancelling was easy: Settings, your name, Subscriptions, FitnessAI, Cancel. Done, and the access runs to the end of the billing period.
A subscription bought on the FitnessAI website is billed by the company, not by Apple, so it does not appear in your iPhone subscription settings at all. FitnessAI confirms both billing channels exist in its own help centre.[15] People who bought that way go looking in the usual place, find nothing, and reasonably conclude they have been trapped. To cancel a website subscription you have to do it inside the app: the gear icon, top right, then Membership, then Cancel.[16]
On Android, cancel through the Play Store: profile icon, Payments and subscriptions, Subscriptions, FitnessAI, Cancel. The help centre gives no Android-specific instructions.
Will You Actually Stick With It?
Probably yes for a month or two if you already lift, and the thing most likely to stop you is not motivation, it is the learning curve. Removing the weekly decision is genuinely powerful, and it is what kept me opening the app. Here is where I think the friction sits.
FitnessAI Pros and Cons
Pros
- Genuinely removes decision fatigue. Open the app, train, close the app
- Prescribes real numbers: sets, reps and a specific load for every exercise
- The animated illustrations shade the working muscles, which teaches you as you go
- The alternatives row is excellent, equipment-matched, and it taught me new movements in three sessions
- You can add exercises on top of the generated session
- The Planning tab is deep: day splits, duration, objective, rep ranges, rest timer and more
- Rest timer is adjustable, with a range you set yourself
- Standings benchmarks you against users in your own age band and bodyweight range
- Upfront about uncertainty: the Learning Phase notice tells you the first suggestion is a guess
- Works properly for home dumbbell training, not just gym training
- Apple Watch app and Apple Health sync
- Cancelling is easy if you subscribe through the App Store
- BodyScan was usable inside my trial without a further charge
Cons
- The injury question at onboarding does not filter your exercises. The company documents this itself[1]
- I declared a knee problem and was offered a jumping squat as a substitution
- The Alternatives list shows what a swap trains but not what it demands. I could find no impact or joint-load information on it
- No named coach, physiotherapist, physician or exercise scientist anywhere in the company
- No guided video for any exercise. Every named competitor has it
- No recovery input. You cannot tell it you slept badly or that this week is harder
- No deload, no warm-up, no cool-down, no rest-day programming
- Opening loads needed correcting. One lift was suggested at 2kg when I train with 8kg
- No guidance on when to progress over weeks, only on correcting today’s number
- Ten simultaneous subscription prices, including three different monthly rates
- No usable free tier. Users report being unable to log a set without paying
- The exercise library size is not published anywhere I could find
- Users report adverts for other apps inside a paid subscription. I saw none in my trial
- The homepage advertises 5 to 30 minute workouts, guided videos and adaptive recovery. I found none of the three
- Challenge points reward inviting friends and posting on social media as heavily as training
- Website subscriptions do not appear in iPhone settings, which is the source of most cancellation complaints
- No nutrition content at all
How FitnessAI Compares to Similar Platforms
Of the five, Caliber is the one I would send a woman over 40 to. All are strength apps rather than class libraries, so it is a fair comparison set. Note the perimenopause row, and note that it reads the same all the way across.
| Feature | FitnessAI | Fitbod | Alpha Progression | Caliber | Ladder |
|---|---|---|---|---|---|
| Our score | 5.5 / 10 | 5.8 / 10 | Not yet reviewed | 7.3 / 10 | 6.9 / 10 |
| Monthly price | $9.99, $14.99 or $19.99 | $12.99 or $15.99 | $12.99 | From $9.00 | From $29.99 |
| Annual price | $59.99 or $89.99 | $79.99 or $95.99 | $79.99 | From $36.00 | From $179.99 |
| Free trial | 3 days on my plan, card required, not on every plan | 7 days, card required, plus 3 free workouts | 14 days, plus a real free tier | Substantial free tier | 7 days, no card required |
| Free tier | Effectively none | 3 free workouts only | Yes, including all 795 exercise videos | Yes, unlimited workout tracking | No |
| Exercise demonstrations | Animated illustrations, no video | Video demonstrations | 795 videos | Video and coach tips | Video plus in-ear audio coaching |
| Library size published | No | Yes, 1,000+ | Yes, 795 | Yes, 800+ | Coach-programmed, not a library |
| Human coaching | None | Trainers by email | None | Yes, 1:1 with a named trainer | Coaches write the programming |
| Named credentials behind it | None found | Trainers on staff | Certified trainers | Elite-level personal trainers | Expert coaches |
| Injury modifications | No. Stated by the company[1] | Manual substitution | Manual substitution | Via your coach | Via your coach |
| Perimenopause content | None | None | None | None | None |
| Nutrition | No | No | No | Yes | Basic AI macro and protein tracking, no dietitian |
| Session length | 20 to 90 min setting, 40 to 60 recommended | Configurable | Configurable | Configurable | Around 30 to 45 min |
| Best suited to | Experienced lifters who want the thinking done for them | Anyone who wants video demonstrations with AI programming | Budget-conscious lifters who want data and a free tier | Anyone who wants a real human coach and nutrition | People motivated by teams and audio coaching |
Competitor prices approximate. Verify on each platform’s website. Scores reflect our independent reviews, see each full review for methodology.
FitnessAI vs Fitbod: Which Is Better for Women Over 40?
For a beginner, Fitbod is the better buy, and it is not close. Both generate AI strength workouts and both adapt to what you log. The difference is that Fitbod shows you video of every one of its published 1,000-plus exercises, and FitnessAI shows you an animated illustration and asks you to read. If you already know how to hinge, brace and press, the illustration is arguably faster and I preferred it. If you do not, the video is the whole point.
On price they are closer than they look. Fitbod is straightforward at $12.99 or $15.99 a month. FitnessAI ranges from $9.99 to $19.99 depending on which ladder you land on, so it may be cheaper or dearer than Fitbod for the same product. Fitbod also gives you seven days to decide against FitnessAI’s three.
Choose Fitbod if you want to be shown the movement. Choose FitnessAI if you want the leanest possible path from opening your phone to finishing a set.
See current FitnessAI pricingAffiliate link, opens fitnessai.com. I earn a commission at no extra cost to you. Full disclosureFitnessAI FAQ
It is a capable strength-training app for any age, but it has nothing built specifically for women over 40. It is a capable strength app at any age, with no recovery input and no coaching. If you already know how to lift it will structure your week well.
No. It asks about injuries at signup, but its own help centre states that the AI does not automatically account for an injury. You have to swap the exercises yourself using the alternatives list, and that list carries no information about impact or load through a joint. I declared a knee problem and was still offered a jumping squat as a substitution.
There is no single price. On the US App Store, ten separate subscriptions run at the same time, including monthly options at $9.99, $14.99 and $19.99, quarterly at $24.99 and $44.99, and annual at $59.99 and $89.99. Check the price on your own paywall, because it varies by user.
There was a 3-day trial on the plan I was shown, and it auto-renews into a paid subscription unless you cancel. The company’s own help centre says some plans do not include a trial period at all, so read the paywall before you tap.
If you subscribed through the App Store, go to Settings, tap your name, tap Subscriptions, select FitnessAI and tap Cancel. It takes seconds. If you subscribed through the FitnessAI website, it will not appear there at all: open the app, tap the gear icon top right, go to Membership and cancel from there. On Android, cancel through the Play Store. Subscribing through the App Store is the safer route.
No, and there is effectively no free tier. Users report being unable to log a set without subscribing. If you want a usable free option, Alpha Progression and Caliber both offer one.
Not for the exercises. Each movement is shown as an animated anatomical illustration with the worked muscles shaded, plus a written Pro Tip and numbered steps. Video appears only in the challenge introductions and the BodyScan capture flow. Fitbod, Ladder, Alpha Progression and Caliber all provide video demonstrations.
It is the wrong choice for most beginners. There is no video, no coaching and nobody checking your form, so you are reading written cues off a phone and deciding for yourself whether a suggested load is sensible. The app asks you to correct its own recommendations, which requires judgement a beginner does not have yet.
Yes, and it works well. You select your equipment at signup and can edit it later in the Planning tab. I tested it with dumbbells and kettlebells at home. Because it prescribes a different load per exercise, adjustable dumbbells make the experience considerably smoother than a fixed set.
Yes. Every exercise opens into an Alternatives row of equipment-matched swaps you can drop in with one tap, and you can add extra exercises on top of the generated session. You can also change the prescribed weight and reps on any set. Swapping is the only way to work around an injury, because the app does not do it for you, and the alternatives list shows you what each swap trains without telling you what it demands of a joint.
Yes. FitnessAI has an Apple Watch app and syncs to Apple Health. I logged from my iPhone throughout and did not test the Watch app across a training block, so I cannot tell you how well it holds up as the primary logging device.
It makes an initial estimate from your signup answers and then adapts based on what you log. The first time it prescribes an exercise it shows a Learning Phase notice saying this is a first recommendation and asking you to adjust the weight and reps as needed. My opening suggestions ranged from reasonable to far too light, including one lift at 2kg when I normally use 8kg per hand.
It estimates body composition from your phone camera using pose estimation, and the company compares its price to a DEXA scan. It is not one. I could use it inside my trial without a further charge, but FitnessAI publishes no price for it and some users report being asked to pay extra on a paid plan, so confirm at signup. Users have also reported readings well out of line with a trainer’s measurement. Treat it as a rough marker to track over time, not a measurement to make decisions on.
It is worth it if you already know how to train, own some dumbbells, and want your weekly structure handled so you can just open the app and go. It is not worth it if you are a beginner, or if you are training around a joint problem and expecting the app to handle it.
Final Verdict
FitnessAI is worth $9.99 a month if you already lift and want your week decided for you, and it is the wrong buy if you are a beginner, training around a joint, or looking for menopause-aware programming. I liked it more than its score suggests, and I want to be straight about why the two do not match. Three days in, I was enjoying it. I opened the app, did what it said, watched my programme while I did it, and came out of it with movements I had not used before. For an experienced woman with dumbbells and no patience for planning her own week, that is a good deal.
The score is what it is because of who this site is for. Five of the twelve capabilities we assess for a midlife body come out at zero, and that is a count rather than a harsh reading. The app takes your age at signup and never uses it again.
What is most likely to sour it is the account rather than the training. Billing and cancellation dominate the recent App Store reviews by a wide margin, and repetitive programming is the next complaint down. Subscribe through the App Store so Apple handles the cancellation, and expect to swap exercises yourself once the rotation starts to repeat.
Use it as what it is: a fast, well-made strength-training engine for someone who already knows how to drive. Pair it with your own cardio and mobility, as I did with a Pilates day and a HIIT day that week. Subscribe through the App Store. And treat every number it gives you as a suggestion, exactly as the company itself tells you to.

Download from the App Store and subscribe there, so Apple handles the billing.
- Price
- $9.99 to $19.99/mo, $59.99 or $89.99/yr
- Free trial
- 3 days, auto-renewing
Sources & Further Reading
- Can the AI take my injury into consideration? FitnessAI Help Centre
- Why was I charged immediately (free trial)? FitnessAI Help Centre
- FitnessAI company profile, Y Combinator
- Fitness AI Gym Workout Planner, App Store listing
- Make the most out of every FitnessAI workout, FitnessAI blog
- FitnessAI launches BodyScan technology for mobile devices, GlobeNewswire
- González-Gálvez N et al. Resistance training effects on healthy postmenopausal women: a systematic review with meta-analysis, Climacteric 2024, PMID 38353251
- Lawford BJ et al. Exercise for osteoarthritis of the knee, Cochrane Database of Systematic Reviews 2024, PMID 39625083
- Schoenfeld BJ et al. Strength and hypertrophy adaptations between low- versus high-load resistance training, Journal of Strength and Conditioning Research 2017, PMID 28834797
- Lopez P et al. Resistance training load effects on muscle hypertrophy and strength gain, Medicine and Science in Sports and Exercise 2021, PMID 33433148
- Plotkin D et al. Progressive overload without progressing load? PeerJ 2022, PMID 36199287
- Greendale GA et al. Changes in body composition and weight during the menopause transition, JCI Insight 2019, PMID 30843880
- Watson SL et al. High-Intensity Resistance and Impact Training Improves Bone Mineral Density and Physical Function in Postmenopausal Women With Osteopenia and Osteoporosis: The LIFTMOR Randomized Controlled Trial, Journal of Bone and Mineral Research 2018, PMID 28975661
- Zhang L et al. A 4-Week Mobile App-Based Telerehabilitation Program vs Conventional In-Person Rehabilitation in Older Adults With Sarcopenia: Randomized Controlled Trial, Journal of Medical Internet Research 2025, PMID 39854716
- Website subscription vs Apple Store subscription, FitnessAI Help Centre
- How can I cancel my subscription? FitnessAI Help Centre
- Do you offer a lifetime membership? FitnessAI Help Centre
- Terms & Conditions, FitnessAI
Related Guides
What To Do Next
Not sure this is the right programme for you?
- → Take our 2-minute quiz for a personalised recommendation.
- → See our best strength training apps for women rankings
- → See our best workouts for women over 40 rankings
- → See how we score every programme using 9 weighted criteria.