Film a set. LiftAI'd extracts 33 body landmarks per frame, counts your reps, measures every joint angle, scores your form against reference patterns, and hands back an annotated video with a coaching report. Same engine family as MoveLens — for the gym.
One phone video, side or front on
Pose estimation — 33 landmarks per frame
Rep segmentation + joint angles (Python)
Form vs reference pattern
Annotated video + report
Each card below is a real video run through the LiftAI'd pipeline. The metrics are what the analysis computes per rep — joint angles, range of motion, tempo, symmetry — plus a form score and a coaching note. The Back Squat card shows real computed output (3 reps, full depth, per-rep scores and the velocity drop); the others are illustrative of the same analysis.
Every angle, rep count, and tempo figure is computed in Python from the pose landmarks — deterministic and reproducible. The AI layer adds the plain-English coaching, never the numbers. LiftAI'd is a training aid for healthy lifters, not a medical or clinical assessment — for workplace and clinical movement risk, that's MoveLens.
One pose-estimation engine, bounded by deterministic maths, applied across domains.
Strength-lifting form coaching for the gym. Rep counting, joint angles, form scoring, annotated coaching video.
You are hereClinical movement & ergonomics to NIOSH and ISO 11228 standards — for physios, OTs and workplace-health teams.
Visit MoveLensEach analysis is a tool an orchestrator can call — score every worker, rank the worksite, draft the plan. See the pattern.
See Live WorkLiftAI'd turns any lifting clip into measured, scored, coachable feedback — the gym-facing surface of the same engine that powers clinical movement analysis.