Before commitment
Will this work for me, or only for the person in the testimonial?
An ecosystem of evidence, support and intelligence
Effort, seen.Evidence, validated.Progress, continues.
An ecosystem of evidence, support and intelligence for members, trainers, clubs, and clinics. What follows is a sample.
Sample · A decision
Evidence qualified- Reliable this session
- Shoulders, yes. Waist, no.
- Comparable to last week
- Shoulders, yes.
- Above the noise floor
- Shoulders, yes.
- Report
- Shoulders moved. Waist withheld.
One decision from a wider set. Weak evidence does not contaminate the rest.
Who it is for · a sample of the set
Members
Progress without false precision.
Trainers
Evidence-linked observations, without reviewing every recording.
Operators
A differentiated experience on ordinary phones.
Partners
A framework that future systems can join, not replace.
More use cases exist. This page shows a few, to give the idea.
The principle
A phone reports only what the evidence can support.
Not another fitness camera, body scanner, or generic scoring app. A disciplined visual measurement system. A video holds useful information about movement, alignment, proportions, and progression. It also holds variation from camera, light, clothing, occlusion, and the model. Most systems treat a recording as usable or unusable, then produce a number. This system asks, for each metric, whether the evidence is there.
Observe
What the camera can actually see in this capture.
Qualify
Which measurements can be computed, and how uncertain each is.
Decide
Report, withhold, defer, or ask for a specific recapture.
How a result is decided
Three questions. Then a report, a withhold, or a recapture.
01
Reliable
How reliable was the measurement in this session?
02
Comparable
How comparable is it with earlier sessions?
03
Enough
Is the difference larger than the expected variation for this person and metric?
A partially obscured video may still support tempo and range, even when it cannot support width or alignment. Early weeks may answer from control, range, load, and consistency. A body-size result appears only when the difference exceeds the validated variation. Relevance alone is not enough.
A sample of the resources
Visual product design
Guided capture, real-time feedback, clear uncertainty.
Computer vision
Pose, segmentation, movement detection, frame selection.
Mobile edge AI
On-device inference, performance, privacy.
Measurement science
Biomechanics, anthropometry, metric definitions.
Statistics and validation
Reliability, personalized noise floors, change detection.
Data and MLOps
Annotation, versioning, testing, monitoring.
Privacy and security
Consent, retention, deletion. Raw media can leave after extraction.
More exists. This is a portion, to give the idea.
A starting point
Evidence, validated. Then the next step.
After commitment
I have been showing up, and I cannot see it. The scale is flat. The mirror is a bad witness.
Planner
The plan. Sets, reps, a progression.
Proof the week happened, on the body that did it.
Scanner · ScanPro
A still. A 3D body, a body-fat number, a report a gym can hand over.
The rep, going beyond the still. Next week still has to change because of the photo.
Form app · FlexifyMe
The pose. A rep score, or a call.
Progress for healthy lifters, across a block, beyond rehab.
Human coach
The rep, in the session. Judgment, and a relationship.
A witness for the days between sessions. The same frame, every week.
Machine · EGYM
The rep, on its machine.
One record across the whole floor, not only the machine that logged it.
Morph
A body. A projection.
Real progress. Never a projected body. The photo stays the photo.
Position
Each tool sees part of the picture. RepsEye is an ecosystem that writes the work into a record, for the person, the trainer, or the club.
ScanPro owns the still. FlexifyMe owns the pose. Mindbody owns the club. EGYM owns the machine. The high-high space, movement and the whole club, is open. RepsEye is built for it.
How a week moves
From the work to the next step.
Capture is guided, and a bad setup is rejected on the phone. The first lifts are a squat pattern, a hinge, a horizontal push, and a horizontal pull. One cue. If the angle cannot support a read, there is no score. Those four moments are one member’s week. The ledger is the only way a week is allowed to change what the next person is told.
01
Intake
- In
- Goal, lifts, and the first frame at the station.
- Out
- The assumptions on screen. Block 1. The first frame is the baseline.
- Changes
- Whether a trial becomes a member.
02
Session
- In
- Phone video of the main lift, after the set.
- Out
- A cue, or a specific recapture. Never a soft score, and never a diagnosis.
- Changes
- Whether today’s work was usable.
03
Weekly proof
- In
- Same frame, same spot. The photo the archive can compare.
- Out
- Report, withhold, or ask for a specific recapture. Only what the evidence supports.
- Changes
- Stay, adjust, or pick up next week.
04
Re-plan
- In
- The proof, the log, and adherence.
- Out
- Next week adapts to what was accepted. The photo stays true.
- Changes
- Renewal, and whether the original promise still holds.
- Not a medical device, not a body-fat lab, and not a diagnosis. A photo read is a coaching signal with a range.
- Not a spotter. Form covers gross, visible faults, and refuses a bad camera.
- Not a filter. An exaggerated morph is out of the product and out of the ads.
- Not a clinician, a physiotherapist, or a dietitian. Escalation is a stop, not a softer plan.
The record
Evidence, validated.
A missed week does not get a kinder image. A flat scale does not get a morph. The sentence has to be something a coach would say out loud.
The scale is flat. The work is not invisible.
Sessions are complete. The overlay shows a shoulder change the member had not named. The sentence is the product. Load is up. The waist photo is not.
- Sessions
- 3 of 4
- Load
- Press up
Load is up. Waist photo is unchanged. Shoulders have moved.
Next block keeps the press and shortens the conditioning.
Sunday lineकंधे आगे बढ़े। कमर नहीं। अगला हफ़्ता पहले से समायोजित है।
Shoulder · MovedWaist · Still
Shoulders moved. waist still.
A registration frame. The regions are named. The photo stays the photo.
Week 12 card
Evidence, validated. The next step follows.
Intake visualization beside the real photo
Behind the bound
The promise and the photo disagree. That disagreement is shown. It is the reason a member can stay through a flat scale without being sold a kinder picture.
The plan changes, or the adherence conversation does. The picture is not improved.
Form
Representative faults. A sample of what can be observed.
A sample of movement metrics. Each clip shows what can be observed, and the signal a camera would use. A metric is reported only when its evidence requirements are met.
17 form · 3 tempo · 3 session · 12 male · 11 female
1
Squat
Lower-body strength and core stability
Knees caving in
FormGoblet squat · Female figure · Front view
- Look for
- Knees drift inside the line of the feet on the way down or on the drive up.
- Camera signal
- Knee-to-knee gap against ankle gap at the bottom; flag when knees cross inside the foot line.
Rounding lower back
FormBack squat · Male figure · Side view
- Look for
- Pelvis tucks under at the bottom and the lower back loses its arch.
- Camera signal
- Hip–lumbar–rib angle through the bottom third of the rep.
Heels lifting
FormBack squat · Female figure · Side view
- Look for
- Heels leave the floor; knees and bar drift ahead of mid-foot.
- Camera signal
- Heel keypoint height and bar offset from mid-foot.
2
Hinge
Posterior-chain strength
Rounded back
FormDeadlift · Male figure · Side view
- Look for
- Spine curves as the bar leaves the floor; hips shoot up before the chest.
- Camera signal
- Curvature of the shoulder–mid-back–hip line during the first pull.
Hyperextending at lockout
FormDeadlift · Female figure · Side view
- Look for
- Leaning back past vertical at the top with ribs flaring.
- Camera signal
- Shoulder behind the ankle vertical at lockout.
Arms instead of hips
FormKettlebell swing · Male figure · Side view
- Look for
- Shoulders lift the bell; hips never fully snap to standing.
- Camera signal
- Hip extension at bell float against arm elevation angle.
3
Push
Upper-body pressing strength
Elbows flaring
FormPush-up · Female figure · Top view
- Look for
- Elbows point straight out to the sides, making a T with the body.
- Camera signal
- Upper-arm-to-torso angle at the bottom; target roughly 30–45°.
Overarching lower back
FormOverhead press · Male figure · Side view
- Look for
- Leaning back to finish the press, ribs popping up.
- Camera signal
- Trunk angle and lumbar curve at lockout.
Poor shoulder mobility
FormOverhead press · Female figure · Side view
- Look for
- Bar finishes in front of the head; arms cannot reach vertical.
- Camera signal
- Shoulder flexion angle at lockout against roughly 170–180°.
Neglecting scapular control
FormPush-up · Male figure · Side view
- Look for
- Shoulder blade peels off the rib cage and the upper back sinks between the shoulders.
- Camera signal
- Scapular border offset from the thoracic line; best seen with a rear camera.
4
Pull
Upper-body pulling strength and posture
Overusing arms
FormBent-over row · Female figure · Side view
- Look for
- The row turns into a curl: the forearm swings up while the elbow stays forward.
- Camera signal
- Elbow travel behind the torso against elbow flexion.
Shrugging shoulders
FormPull-up · Male figure · Front view
- Look for
- Shoulders ride up toward the ears and the neck disappears.
- Camera signal
- Ear-to-shoulder distance through the pull.
Rounding upper back
FormBent-over row · Female figure · Side view
- Look for
- Chest caves and the upper back hunches over the weight.
- Camera signal
- Thoracic curvature and head-to-trunk angle.
Jerking movements
TempoPull-up · Male figure · Side view
- Look for
- Swinging and kicking to get up, then dropping fast.
- Camera signal
- Velocity spikes in shoulder height and body-swing amplitude.
5
Core stability
Anti-extension and anti-rotation
Sagging hips
FormForearm plank · Female figure · Side view
- Look for
- Hips drop below the line from shoulder to ankle as the hold goes on.
- Camera signal
- Hip distance below the ankle–shoulder line over the hold.
Overarching lower back
FormDead bug · Male figure · Side view
- Look for
- Low back peels off the floor as the leg reaches long.
- Camera signal
- Gap between lumbar spine and floor while limbs extend.
Holding breath
TempoForearm plank · Male figure · Side view
- Look for
- No breathing rhythm; strain builds and the body starts to shake.
- Camera signal
- Rib-cage motion amplitude and tremor frequency.
6
Locomotion and conditioning
Cardio and metabolic fitness
Skipping warm-up
SessionRun start · Female figure · Side view
- Look for
- Going from still to full effort with no gradual ramp.
- Camera signal
- Session timeline: no low-intensity block before peak cadence. One rep cannot show this.
Overstriding and heel striking
FormRunning · Male figure · Side view
- Look for
- Foot lands well ahead of the hip, heel first, with a straight knee.
- Camera signal
- Ankle ahead of hip at contact and foot angle at touchdown.
7
Mobility and flexibility
Range of motion and joint health
Skipping mobility work
SessionDeep squat test · Male figure · Side view
- Look for
- Depth stops short; heels rise and the chest dumps forward to compensate.
- Camera signal
- Depth and ankle/hip range trend across sessions, not one rep.
Overstretching cold muscles
SessionStanding forward fold · Female figure · Side view
- Look for
- Bouncing to force range before the muscles are warm.
- Camera signal
- Oscillation in fold depth, and a stretch logged before any warm-up in the session.
8
Balance and function
Stability and proprioception
Rushing movements
TempoSingle-leg RDL · Female figure · Side view
- Look for
- Drops fast, overshoots, then wobbles and taps the free foot down.
- Camera signal
- Rep tempo and trunk sway frequency.
Poor foot stability
FormSingle-leg stance · Male figure · Front view
- Look for
- Arch collapses, ankle rolls in, and the knee and hip wobble over the foot.
- Camera signal
- Ankle lateral sway and knee-over-foot alignment.
Form faults show inside a single rep. Tempo faults only show over time, in the signal strip. Session faults are about what a person does or skips across a workout, so video alone cannot prove them. Figures are schematic; angles are exaggerated so each fault reads at a glance.
The Indian context
Evidence that fits the body it is for.
Strong, athletic, and pleasing are allowed only inside what lasts and what does not hurt. The order is fixed. Lasting, then strong, then proportion. Pain ends the score. The rate caps the picture. These figures stay chart v1 until a person signs a later version. That argument is the ledger, not this calculator.
The Indian risk sits at the waist, often while a Western chart still calls the BMI normal.
BMI band
Asian Indian adults
18.5–22.9 narrower normal. 23–24.9 overweight. ≥25 obesity. Under 18.5 is not a leanness goal.
Sets whether a block may talk about a smaller waist at all. Under 18.5, the block feeds the work. It does not shrink the picture.
Does not stage disease, and does not use the Western 25 / 30 chart.
Misra et al., Journal of the Association of Physicians of India, 2009;57:163–170. The 2025 India Obesity Commission starts increased adiposity above BMI 23 and does not ask RepsEye to diagnose stage 2.
Waist circumference
Women and men, measured at the abdomen
Action level: women ≥80 cm, men ≥90 cm.
The first regional note when the member is outside it. Shoulders can be named only after this is faced.
Will not quote the Western action levels of 88 cm and 102 cm. Those are the wrong body.
Misra et al., JAPI, 2009. The same waist levels are the ones the Indian discussion still uses.
Waist-to-height
Any adult, using their own height
Keep the waist under half the height. Bound: 0.50.
The personal number. A 5′2″ member and a 6′ member do not share a waist target. The projection may not draw a waist smaller than this bound unless the photo has already gone there.
Not a tailor’s ideal. Not a reason to keep cutting once the member is under it.
Ashwell M, Gibson S. BMJ Open. 2016;6:e010159. The 2025 Indian definition includes waist-to-height alongside waist; RepsEye uses 0.50 and does not invent a tighter private cutoff.
Waist-to-hip
Women and men
Raised-risk action: women ≥0.85, men ≥0.90.
Secondary. A larger hip can hide a high waist inside this ratio, so the waist and the height come first.
Not an hourglass target.
World Health Organization. Waist circumference and waist–hip ratio: report of a WHO expert consultation. Geneva, 2008.
Grip strength floor
Older Asian adults, as a floor
Low strength: women under 18 kg, men under 28 kg.
Lasting muscle, not a Sunday physique score. For the 22–40 beachhead it is a floor to know. It is not the weekly card.
Not a test RepsEye infers from a photo.
Chen et al. Asian Working Group for Sarcopenia 2019. Journal of the American Medical Directors Association. 2020;21(3):300–307.
Calf circumference
Older Asian adults, screening only
Screen for low muscle: women under 33 cm, men under 34 cm.
A floor for muscle that lasts. It is not a calf size to build toward.
Will not turn a sarcopenia screen into an aesthetic target.
Chen et al., AWGS 2019, JAMDA 2020.
How it learns
What is followed teaches what comes next.
Two feeds. Member weeks, as counts. Research, as a slip a person can sign. Neither feed is a photo. Neither feed is allowed to improve a picture already issued. This desk is a worked example. It forgets when you leave.
01
Freeze
The path is named under a chart version: the dose, the assumption, and what it refused. The check uses the card that was issued, not a later opinion.
02
Count
A followed week has the sessions at that dose and a Sunday frame the camera accepted. A miss, a pain stop, and a bad camera are stored, then excluded.
03
Check
Each followed week is ahead, inside, or behind that original card. Eight followed weeks, and at least six in ten of them the same way, before a change can even be proposed.
04
Sign
A person signs the next chart, or refuses the slip. Lasting lines do not move on member weeks. Issued cards are not recomputed. Photos never enter.
- Chart
- v1
- Lasting
- Frozen
- 80 / 90 cm · 0.50 · BMI 23
- Next press aim
- As charted
- New intakes only
- Next fat-loss cap
- 0.5% a week
- Never faster from this desk
Issued cards stay on the version they were promised. Women 80 cm, men 90 cm. Waist under half the height. BMI floor 18.5, band from 23, obesity from 25. No body-fat percentage.
Paths issued under chart v1
File a week on this path
Novice horizontal push
6 of 8 followed weeks landed behind the novice push aim. The next chart may lower that aim one step. It cannot move the waist lines, and it cannot redraw a shoulder.
Research slips
Unsourced lookbook
Unsigned
Waist under 0.45 of height reads as more athletic.
A tighter ratio with no source is a look. The only signature is a refusal.
A US product chart
Unsigned
Use 88 cm for women and 102 cm for men so the app matches an international tool.
A Western action level is not a correction for this body.
India Obesity Commission, 2025
Unsigned
Increased adiposity begins above BMI 23. The commission does not ask this product to diagnose a stage.
The band is already on the card. Signing adds the citation. It does not add a diagnosis.
Physique-coaching rate note
Unsigned
When adherence is new, cap a fat-loss picture at about 0.25% of body mass a week.
It can be signed only after the fat-loss path earns it, or held until then. It cannot speed a picture up.
What this sitting changed
Nothing filed yet. The opening read is already a check: the recomp path held, the novice press is ready to sign down one step, and the waist lines are not open for negotiation.
- Photos do not enter the learning set. The archive stays a proof, not a training pile.
- Names and centimetres do not enter the cohort file. The gap does: ahead, inside, or behind.
- A testimonial body is not a source. A Western chart is not a correction.
- An unfollowed path cannot acquit a recommendation, and it cannot convict one.
Sources on this chart
- Misra et al., JAPI, 2009. Waist 80 / 90 cm. BMI band from 23. Floor at 18.5.
- Ashwell and Gibson, BMJ Open, 2016. Waist under half the height.
- Strength bands are coaching convention, not an Indian survey. No elite column.
Who it serves
One ecosystem. Many roles.
Members first, in one city, on one goal. Trainers next, because the close is the case study. Gyms after that, with a holdout. Employers last, and blind to body images by default.
Meera, 29, Bengaluru
See whether the work is landing, on your body, every week.
Trains three times a week. Has quit two apps and one trainer. Cannot tell if her shoulders have changed.
- Offer
- The app, and a weekly proof she can trust.
- Price shape
- ₹499–999 a month after the trial. Working price ₹699.
- How it starts
- A goal, three guided photos, a projection with the assumptions on it, and block 1. She pays after the first proof.
- Do not say
- Dream body. Guaranteed transformation. Medical-grade fat.
How it is offered
Support that follows the evidence.
If a member hesitates to take the Sunday photo because it costs something, the product is already broken. Meter extra projection scenarios if a meter is required. Do not meter the week.
Beachhead
Direct
The member
₹699
a month, working price
- Range for the beachhead is ₹499–999 after the trial.
- Android first. English first. Simple Hindi on the Sunday line.
- Bengaluru, and one other metro. Recomposition. People who already train.
- The weekly proof is not metered. A Sunday photo is the reason to stay.
Next
Trainer seat
The trainer or the studio
Per client
plus a platform fee
- Same capture as the member app. Their note sits beside the image.
- Judged on active clients per trainer, not on consumer ads.
- They show the projection unedited. If they have to fix it, the channel has failed.
- Morph creatives are disallowed. They select the wrong user.
Pilot
Gym layer
The operator
Per site
or per active member
- Phone and QR. No hardware dependency.
- Kill-criterion agreed before the pilot starts.
- Watched on trial-to-member and day-30 return, against a holdout.
- Two or three sites. Not a rollout dressed up as a test.
Corporate wellness is a later door: per employee per year, adherence only, no body images unless the employee opts in. No revenue target is set for year one. It is a proof year.
Lines the brand will not say
To a member
Dream body. Guaranteed transformation. Medical-grade fat.
To a trainer
The app replaces you. The picture is the result.
To a gym
Footfall, dwell, or shares are how we will know it worked.
To an employer
Company-level body-composition outcomes.
How it is held
The record is kept. The work continues.
Private by default
What is stored, for how long, and who at a gym or an employer can ever see it is said at capture. The default is no one.
Bounds are versioned
The rate a projection may show is a rule, not a prompt. Marketing cannot outrun it. A later chart does not retouch a card already issued. Photos are not how the chart learns.
The card is immutable
Once the week-12 card is generated, it is not retouched. Public case studies are opt-in. There is no public transformation feed.
Eighteen, and a stop
Age gate at 18. No minor body imagery. No prompts aimed at extreme leanness, or at someone else’s body. Distress routes to a stop-and-refer screen.
How it begins
How a partnership begins.
The decision to continue is week-4 paid retention against a plan-only control, and whether people actually take the Sunday photo.
- 01
Weeks 0–2
Capture protocol, the Indian chart, and the ledger floor. Eight followed weeks on a path before that path may move a chart. Reviewed with two trainers.
- 02
Weeks 3–6
Closed beta, about fifty members. Weekly proof completion is the only north star.
- 03
Weeks 7–12
Paid trial. Week-4 retention against a plan-only control in the same app.
- 04
Then
Continue, narrow, or stop. On proof completion and week-4 paid retention. Not on downloads.
A conversation, not a waitlist
Write the brief before the meeting.
Same facts, different job. Pick who is in the room. The brief carries the line they can repeat, and the line this brand will not say. It stays on this device until you copy it.
RepsEye
Progress, continues.
Every record we show is one the member made. Intake on one side. Week 12 on the other. Next week comes from that pair.