Most free workout generators hand you a plan and never tell you where it came from. You get a split, some sets and reps, and no way to judge whether any of it fits you. This one works the other way around. Every rule it applies is published on this page, with the paper behind it linked at first mention, so you can audit the plan or change it on purpose rather than by feel.
It is also honest about what it is. The generator below is a rule engine, not a language model and not a coach. It reads five answers and applies a fixed decision tree. That is enough to produce a genuinely good starting plan, because the decisions that matter most (how often to train each muscle, how many hard sets to do, what rep range to use, how close to failure to go) are well described in the research. What it cannot do is watch you train.
How this generator chooses your plan
Five inputs, five decisions. Nothing here is invented, and nothing is randomized.
- Days per week picks the split. Two or three days becomes full body, four becomes upper and lower, five becomes push, pull, and legs plus an upper and a lower day. Every option trains each muscle group at least twice a week, because Schoenfeld et al. 2016 in Sports Medicine found that training a muscle twice weekly produced more hypertrophy than once weekly when total volume was equal. Grgic et al. 2018 found the same pattern for strength. This is why a 3-day plan here is full body rather than a bro split.
- Experience picks sets per exercise. Beginners get 2 sets, which sits inside the 1 to 3 sets per exercise the ACSM position stand on progression models (Ratamess et al. 2009) recommends for novice lifters. Intermediate and advanced lifters get 3 (2 on the consistency goal, which is deliberately built around the minimum effective dose), and advanced lifters get one extra set on the first exercise of each session. Weekly volume then rises with training age, which matches the dose-response curve Schoenfeld et al. 2017 mapped in the Journal of Sports Sciences, where muscle growth kept climbing with weekly set count.
- Goal picks the rep range, rest, and effort. Building muscle uses 6 to 10 reps on compounds with 90 to 180 seconds of rest. Losing weight uses 10 to 15 reps with 45 to 75 seconds and adds a conditioning finisher, which is dropped on a 10 to 20 minute session because it genuinely will not fit. Getting consistent uses 8 to 12 reps with 60 to 90 seconds and fewer exercises. The ranges come from the repetition continuum re-examined by Schoenfeld et al. 2021 in Sports, and the longer rest for muscle building comes from Singer et al. 2024, whose Bayesian meta-analysis favored rest periods over 60 seconds for hypertrophy.
- Equipment picks the exercise for each movement slot. Every session is built from movement patterns first (squat, hinge, horizontal push, horizontal pull, vertical push, core) and then filled with the best available exercise for your equipment. On bodyweight the rep ranges widen, because Schoenfeld et al. 2017 in the Journal of Strength and Conditioning Research found similar hypertrophy from low and high loads when sets are taken close to failure.
- Session length is a real time budget, not a label. The generator budgets roughly 40 seconds of work plus your prescribed rest for every set, adds a warm-up, and then trims exercises until the session actually fits. A 20-minute session with 2-minute rests genuinely cannot hold 6 exercises, so it does not pretend otherwise. The time-efficiency logic follows Iversen et al. 2021 in Sports Medicine.
Can AI create a workout plan?
Yes, and the honest answer is that the structural part is not very hard. Choosing a split, assigning sets and reps, ordering compounds before isolation, and matching exercises to available equipment are rule-following tasks. A decision tree does them well. A language model does them well too, most of the time, which is why asking ChatGPT for a workout plan often produces something reasonable.
The gap opens after the first session. A generated plan is a snapshot of what was sensible before you trained. It does not know that you finished all 3 sets of 10 with room to spare, that your left shoulder pinched on the overhead press, that you slept 5 hours, or that you skipped last Thursday and Friday. Those are the inputs that change what the next session should be, and a form on a web page never sees them.
That is the real dividing line between a workout generator and an AI personal trainer app. A generator produces a plan. A coach produces the next session, given the last one. FitCraft's coach Ty is the second kind: it demonstrates every movement in 3D so you can see the form before you attempt it, records what you actually completed, and reshapes the following week around it. We wrote up the mechanics in how AI adapts workouts and what an AI fitness coach actually does. The generator on this page is deliberately the simpler thing, and it is free with no signup.
How days per week picks your split
Frequency is the input people most often get wrong, usually by picking a split that trains each muscle once a week. Here is what this generator does instead.
| Days | Split | Session names | Each muscle trained |
|---|---|---|---|
| 2 days | Full body | Full body A, Full body B | 2 times per week |
| 3 days | Full body | Full body A, B, C | 3 times per week |
| 4 days | Upper / Lower | Upper A, Lower A, Upper B, Lower B | 2 times per week |
| 5 days | PPL + Upper/Lower | Push, Pull, Legs, Upper, Lower | 2 times per week |
Notice what is missing. There is no one-body-part-per-day split at any frequency, because at 3 to 5 training days that structure cannot hit each muscle twice. Push, pull, and legs only appears at 5 days, where two extra full upper and lower sessions bring every muscle back to twice-weekly. The evidence for that choice is covered in our review of training frequency research.
Reps, rest, and effort by goal
The three goals are not three different sciences. They are three different points on the same continuum, with different rest periods and a different amount of conditioning attached.
| Goal | Reps (loaded) | Reps (bodyweight) | Rest | Effort target |
|---|---|---|---|---|
| Build muscle | 6 to 10 | 8 to 15 | 90 to 180 s | 0 to 3 reps in reserve |
| Lose weight | 10 to 15 | 12 to 20 | 45 to 75 s | 1 to 3 reps in reserve |
| Get consistent | 8 to 12 | 10 to 15 | 60 to 90 s | 2 to 4 reps in reserve |
Accessory and core work always runs at higher reps than the compounds in the same session, typically 10 to 15 loaded or 12 to 20 on bodyweight, and timed holds like planks are prescribed in seconds rather than reps.
Reps in reserve is how many more reps you could have completed when you racked the weight. Two reps in reserve means you stopped with two left. The generator moves that target with experience, because Refalo et al. 2023 found that hypertrophy is fairly forgiving of stopping a few reps shy of failure, while the injury and fatigue cost of training to failure is not. Beginners get the widest margin. We go deeper in training to failure versus reps in reserve.
Rest periods matter more than most plans admit. Cutting rest to 30 seconds to "save time" reliably reduces the reps you can complete on the next set, which lowers the volume that actually drives growth. Our summary of the evidence is in rest periods and muscle growth.
Worked examples
Seven input combinations and exactly what the generator returns for each, so you can sanity-check it against your own situation without touching the tool.
| Inputs | Split | Per session | Sets per week | Reps | Est. time |
|---|---|---|---|---|---|
| Muscle, beginner, 3 days, dumbbells, 30 to 45 min | Full body ×3 | 6 exercises, 2 sets | 36 | 6 to 10 | 34 min |
| Muscle, intermediate, 4 days, full gym, 30 to 45 min | Upper / Lower | 5 exercises, 3 sets | 60 | 6 to 10 | 41 min |
| Muscle, advanced, 5 days, full gym, 30 to 45 min | PPL + Upper/Lower | 5 exercises, 3 sets plus 1 | 80 | 6 to 10 | 44 min |
| Muscle, beginner, 3 days, bodyweight, 10 to 20 min | Full body ×3 | 3 exercises, 2 sets | 18 | 8 to 15 | 18 min |
| Weight loss, beginner, 3 days, bodyweight, 20 to 30 min | Full body ×3 | 5 exercises, 2 sets, plus finisher | 30 | 12 to 20 | 27 min |
| Weight loss, intermediate, 4 days, dumbbells, 30 to 45 min | Upper / Lower | 5 exercises, 3 sets, plus finisher | 60 | 10 to 15 | 35 min |
| Consistency, beginner, 2 days, bodyweight, 10 to 20 min | Full body ×2 | 3 exercises, 2 sets | 12 | 10 to 15 | 15 min |
The bottom row is the one worth staring at. Twelve working sets a week across two 15-minute sessions looks like almost nothing, and it still clears the two-day muscle-strengthening floor in the Physical Activity Guidelines for Americans (Piercy et al. 2018). Androulakis-Korakakis et al. 2020 found that even trained lifters gained meaningful strength from remarkably small weekly doses. If the choice is between a perfect plan you abandon and a small plan you keep, the small plan wins every time. More on that in the minimum effective dose of exercise.
You have a plan now. The next problem is doing it 40 more times.
FitCraft's coach Ty is a 3D personal trainer that demonstrates every movement, tracks what you actually finish, and rebuilds next week around it. Take the free 2-minute assessment to see your starting plan.
Take the Free Assessment Free • 2 minutes • Personalized to youProgressive overload is the rule that makes any plan work
A generated plan is a container. Progressive overload is what you put in it. If week 8 looks exactly like week 1, the plan did nothing, no matter how well designed the split was.
The rule this generator gives you is deliberately boring. Work in the prescribed rep range. When you hit the top of the range on every set of an exercise, add load next session and drop back to the bottom of the range. Repeat. On dumbbells that is usually 5 lb (about 2.5 kg) per jump, in a full gym 5 to 10 lb, and on bodyweight there is no load to add, so you progress by adding reps and then moving to a harder variation: incline push-ups to kneeling push-ups to floor push-ups to close-grip push-ups.
Two things people do instead, both of which stall progress:
- Changing exercises every week. Variety feels like progress and destroys your only measurement. If the movement changes, you cannot tell whether you got stronger. The idea that muscles need to be "confused" does not survive contact with the evidence, which we cover in the muscle confusion myth.
- Adding weight before the reps are there. Jumping load while your top set is still at the bottom of the range means every set gets shorter and weekly volume quietly falls. Reps first, then load. Our beginner walkthrough is in progressive overload for beginners.
Run the same plan for 4 to 8 weeks. If a lift stalls for two or three weeks in a row while your sleep and food are steady, take a lighter week before you take a new program. The case for that is in the science of the deload week.
What changes with your equipment, and what does not
The movement patterns never change. Every session in this generator is built from the same skeleton: a knee-dominant leg exercise, a hip-dominant leg exercise, a horizontal push, a horizontal pull, a vertical push, and core work. Only the exercise that fills each slot changes.
No equipment. Squats, glute bridges, push-ups, inverted rows, pike push-ups, and planks cover every slot. The rep ranges widen because you cannot add load, and progression runs through harder variations instead. That is a real plan, not a compromise, and the results are documented in our review of bodyweight training results. If you train this way, our home workouts guide, our roundup of the best bodyweight exercises, and FitCraft's home workout app with no equipment are the natural next stops.
Dumbbells. One adjustable pair unlocks nearly everything: goblet squats, Romanian deadlifts, chest presses, rows, shoulder presses, and every accessory. This is the best value-per-square-foot setup in home training, and it is what FitCraft's dumbbell workout app is built around. Start lighter than you think and let the progression rule do the work.
Full gym. Barbells and machines mainly buy you smaller load jumps and more stability at heavy weights, which matters most for strength. The exercises the generator selects are the same patterns, so load them with whatever version your gym has. Every link in the plan goes to a FitCraft form guide for that movement.
When a static plan stops being enough
A generated plan is right on the day you generate it and slowly gets less right afterward. Three specific things break it.
Progression outruns the prescription
Eight weeks in, the loads on the page are wrong for you, and the volume that was right when you were untrained is now the volume you have adapted to. Someone or something has to decide what the next block looks like. A static plan cannot, because it does not know what you actually completed.
Plateaus need a diagnosis, not a new template
When a lift stalls, the fix depends on the cause: too little recovery, too little food, too much volume, technique breaking down under load, or simply not enough time on the same movement. A generator that only knows your goal and your equipment cannot tell those apart. It will hand you a different template, which is the one intervention that does not address any of them.
Adherence, which is the part that actually decides your results
This is the real one. The plan you follow beats the plan that was theoretically optimal, and most people do not stop training because their split was wrong. They stop because the week got hard, they missed two sessions, and the plan on the page still says Week 3 Day 1 while their life has moved on. Our summary of workout adherence statistics is bleak reading. A plan that never adjusts to a missed week quietly turns into a document that reminds you you are behind.
This is where FitCraft is a different product, not a better version of this page. It watches what you finish, adapts the next session accordingly, teaches technique with 3D demonstrations instead of assuming you already know it, and wraps the whole thing in a progression system built to keep you coming back. That last part is deliberate: the gamified workout app structure and the motivation mechanics exist because adherence, not programming, is the constraint for most people. If you are starting from zero, the beginner path is the place to start. There is a free version, and premium runs $19.99 per month or $119.99 per year (50 percent savings) with a 7-day free trial.
When to ignore this generator
The plans here are built for healthy adults with no current injury. A few situations call for professional programming instead of a general template.
- Current injury or pain during a movement. Pain is information, not something to train through. See a physiotherapist or sports medicine clinician for an assessment and movement-specific modifications.
- Cardiovascular disease, uncontrolled high blood pressure, or recent cardiac symptoms. Resistance training is usually beneficial here, but the prescription should come from a clinician or a clinical exercise physiologist, not a web form.
- Pregnancy and the postpartum period. Load tolerance, core work, and appropriate intensity all change across pregnancy and recovery. Follow your obstetric care team, and see our postpartum fitness guide for context.
- Eating disorders or recovery from one. Exercise prescriptions tied to weight loss can be harmful. Work with a treatment team before adding structured training volume.
- Osteoporosis, recent surgery, joint replacement, or a hernia. Certain loading patterns are contraindicated. Get movement-specific clearance first.
- Children and adolescents. The set and rep prescriptions here are drawn from adult research. Youth resistance training is safe when supervised, but it should be programmed for the age group.
If you have joint pain rather than an acute injury, our training with joint pain guide covers modifications that usually work. If you are over 60 or returning after a long break, fitness over 60 and starting from zero fitness are better starting points than any generated split.
Related reading
References
- Ratamess NA, Alvar BA, Evetoch TK, et al. "American College of Sports Medicine position stand. Progression models in resistance training for healthy adults." Medicine and Science in Sports and Exercise. 2009;41(3):687-708. doi:10.1249/MSS.0b013e3181915670
- Schoenfeld BJ, Ogborn D, Krieger JW. "Effects of Resistance Training Frequency on Measures of Muscle Hypertrophy: A Systematic Review and Meta-Analysis." Sports Medicine. 2016;46(11):1689-1697. doi:10.1007/s40279-016-0543-8
- Schoenfeld BJ, Ogborn D, Krieger JW. "Dose-response relationship between weekly resistance training volume and increases in muscle mass: A systematic review and meta-analysis." Journal of Sports Sciences. 2017;35(11):1073-1082. doi:10.1080/02640414.2016.1210197
- Schoenfeld BJ, Grgic J, Ogborn D, Krieger JW. "Strength and Hypertrophy Adaptations Between Low- vs. High-Load Resistance Training: A Systematic Review and Meta-analysis." Journal of Strength and Conditioning Research. 2017;31(12):3508-3523. doi:10.1519/JSC.0000000000002200
- Schoenfeld BJ, Grgic J, Van Every DW, Plotkin DL. "Loading Recommendations for Muscle Strength, Hypertrophy, and Local Endurance: A Re-Examination of the Repetition Continuum." Sports. 2021;9(2):32. doi:10.3390/sports9020032
- Grgic J, Schoenfeld BJ, Davies TB, Lazinica B, Krieger JW, Pedisic Z. "Effect of Resistance Training Frequency on Gains in Muscular Strength: A Systematic Review and Meta-Analysis." Sports Medicine. 2018;48(5):1207-1220. doi:10.1007/s40279-018-0872-x
- Refalo MC, Helms ER, Trexler ET, Hamilton DL, Fyfe JJ. "Influence of Resistance Training Proximity-to-Failure on Skeletal Muscle Hypertrophy: A Systematic Review with Meta-analysis." Sports Medicine. 2023;53(3):649-665. doi:10.1007/s40279-022-01784-y
- Androulakis-Korakakis P, Fisher JP, Steele J. "The Minimum Effective Training Dose Required to Increase 1RM Strength in Resistance-Trained Men: A Systematic Review and Meta-Analysis." Sports Medicine. 2020;50(4):751-765. doi:10.1007/s40279-019-01236-0
- Iversen VM, Norum M, Schoenfeld BJ, Fimland MS. "No Time to Lift? Designing Time-Efficient Training Programs for Strength and Hypertrophy: A Narrative Review." Sports Medicine. 2021;51(10):2079-2095. doi:10.1007/s40279-021-01490-1
- Singer A, Wolf M, Generoso L, et al. "Give it a rest: a systematic review with Bayesian meta-analysis on the effect of inter-set rest interval duration on muscle hypertrophy." Frontiers in Sports and Active Living. 2024;6:1429789. doi:10.3389/fspor.2024.1429789
- Wewege MA, Desai I, Honey C, et al. "The Effect of Resistance Training in Healthy Adults on Body Fat Percentage, Fat Mass and Visceral Fat: A Systematic Review and Meta-Analysis." Sports Medicine. 2022;52(2):287-300. doi:10.1007/s40279-021-01562-2
- Piercy KL, Troiano RP, Ballard RM, et al. "The Physical Activity Guidelines for Americans." JAMA. 2018;320(19):2020-2028. doi:10.1001/jama.2018.14854
How this generator differs from typical AI workout generators
Most free workout generators do one of two things. Either they output a plan with no methodology at all, so you cannot tell whether the split, the set count, or the rep range came from research or from a template someone typed in 2011, or they wrap a language model that produces plausible-looking programming that changes every time you ask. Both approaches share the same weakness: the plan is unauditable. You cannot check the reasoning, so you cannot correct it when it is wrong for you.
This generator publishes its entire decision tree on the page and links the primary source behind every rule. It is deterministic, so the same inputs always return the same plan and you can compare two configurations honestly. It treats session length as a genuine time budget and trims exercises rather than prescribing a 6-exercise workout that cannot fit in 20 minutes with 2-minute rest periods. It refuses to produce a split that trains a muscle once a week, at any frequency, because the frequency literature does not support it. Every exercise it selects links to a written form guide rather than assuming you already know the movement. And it states plainly that it is a rule engine rather than implying that a form submission constitutes coaching, because the thing that actually changes outcomes is what happens after session one, and no static generator, this one included, can do that part.
Frequently Asked Questions
What is the best free AI workout generator?
The best free workout generator is the one that shows its work. Most free tools output a plan with no explanation of where the set counts, rep ranges, or split came from, which makes the plan impossible to audit or adjust. This generator publishes every rule it uses and cites the paper behind it: split frequency from Schoenfeld et al. 2016, weekly volume from Schoenfeld et al. 2017, rep ranges from Schoenfeld et al. 2021, novice set counts from the ACSM position stand, and time-efficient session structure from Iversen et al. 2021. It is free, needs no signup, runs entirely in your browser, and links every exercise to a form guide.
Can AI create a workout plan?
It can create a reasonable starting plan. A language model or a rule engine can pick a split, assign sets and reps, and choose exercises that fit your equipment, because those decisions come from published training principles. What software cannot do from a single form is see how the last session actually went. A plan only becomes coaching when something reads your completed sets, your recovery, and your missed days and changes the next session in response. That is the difference between the generator on this page, which produces the same plan every time you enter the same answers, and an app that adapts session to session.
How many days a week should I work out?
Two to four days of resistance training covers almost everyone. The Physical Activity Guidelines for Americans set the floor at two muscle-strengthening days per week. Schoenfeld et al. 2016 found that training a muscle twice a week beat once a week for hypertrophy when weekly volume was equal, which is why this generator uses full body plans at two and three days rather than a split that hits each muscle once. Beyond four days, added benefit comes mostly from spreading the same weekly volume across more sessions rather than from doing more work overall.
Is an AI workout plan as good as a personal trainer?
For the structure, it is close. Split selection, set and rep prescription, and exercise ordering are well-described in the literature, and software applies them consistently. For everything else, it is not close. A trainer watches your technique, catches a compensation before it becomes an injury, changes the plan when your shoulder hurts, and notices when you stop showing up. Software can only match that if it collects the same information. Look for a tool that demonstrates form, logs what you actually completed, and changes the next session based on it rather than handing you a static PDF.
How many sets and reps should a beginner do?
Two sets per exercise, 6 to 12 reps, on 4 to 6 exercises per session is plenty to start. The ACSM position stand (Ratamess et al. 2009) recommends 1 to 3 sets per exercise for novice lifters, and Androulakis-Korakakis et al. 2020 found meaningful strength gains from surprisingly small doses in trained lifters, so an untrained lifter needs less than most plans prescribe. Stop each set with 2 to 3 reps still in reserve. The goal in the first months is learning the movement and building the habit, not accumulating maximum volume.
Can I get a good workout plan with no equipment?
Yes, with one adjustment. Schoenfeld et al. 2017 found that low-load and high-load training produce similar muscle growth when sets are taken close to failure, so bodyweight training works as long as the sets are hard. Because you cannot add load, this generator raises the bodyweight rep ranges (8 to 15 on compounds instead of 6 to 10) and progresses you by adding reps first, then moving to a harder variation such as incline push-ups to floor push-ups to close-grip push-ups. Strength gains at very heavy loads are the one thing bodyweight training genuinely cannot match.
How often should I change my workout plan?
Less often than most people do. Run the same plan for at least 4 to 8 weeks so you can actually add reps or load to the same exercises and see whether it is working. Changing exercises every week feels productive but removes the only signal that tells you whether you are progressing. Change the plan when progress genuinely stalls for two to three weeks in a row, when your schedule or equipment changes, or when something hurts. Novelty for its own sake is not a training variable.