How we check every pose
Repglyph poses are validated against cited professional and clinical standards. This page explains what that means, what the automated checks cover, and what has not been done yet.
- Cited references
- 273
- Automated checks per run
- 2,633
- Movement families verified
- 53/54
Status
The checks are automated measurements against documented, cited thresholds. They are not a clinical assessment. The clinical form report that collects the sources and results is marked “Draft for clinical review”, and independent clinician review is pending.
The written standard
Before a pose ships, it is measured against our written form standard, which draws on 273 references: ACE, NSCA, ACSM and NASM coaching guidance, AAOS and CDC joint-range norms, and peer-reviewed biomechanics. Every threshold names its source. Where sources disagree, both values are recorded and the one we chose is marked.
The automated checks
An automated suite samples every movement, variant and body type at 32 points in every rep. It measures joint angles, contacts, balance, grip width and range of motion. Each check passes, fails or is exempt with a written reason. The current run is 2,633 checks across 183 movement and variant rows on three bodies, with 0 failures.
53 of the 54 movement families are verified and 1 is partial. Where we couldn't find a source, we say so: no peer-reviewed joint angles exist yet for the carry family's sled push and pull, so that family is still marked partial.
Example checks
ROM · hip extension≤ 20°
ROM · knee flexion≤ 145° active
wrist.rangeExtension ≤ 70° active. About 90° only for flat-palm floor support. ≤ 15-20° when holding an implement.
squat.shinForward shin tilt 15-40° with the heels flat
pushup.handsWrist within 3 cm of directly under the shoulder
Each tag names a source for the threshold. “house” marks a threshold from our own written standard.
Checked against people
We also compare the figures with joint angles from licensed motion-capture datasets: CMU, UI-PRMD and the ARKit-versus-Vicon set from Reimer et al., 2022. Any keyframe that drifts outside the recorded range is corrected.
What this is not
The figures illustrate exercise technique. They aren't fitted to any individual, and they don't replace a coach, physiotherapist or doctor.