Case study — Formax
AI form analysis for lifting videos
Record or upload a lifting video and Formax scores the movement on form, control, range of motion and intensity — then pins its feedback to the exact moments it happened, on an interactive timeline you can scrub frame by frame.
A subscription product I took from idea to the App Store alone: the vision pipeline, video processing and storage, the mobile experience, the backend, billing, privacy review, and every release since.
What shipped
Stack
Engineering
01
Workout clips are large, shot in poor gym lighting, and arrive from arbitrary devices. The pipeline compresses on-device before upload, stores originals in object storage rather than a database, and processes asynchronously so the app never blocks on analysis.
02
Feedback is only useful if it maps to a moment. The model is constrained to emit scored segments with start and end timestamps, which the client renders directly onto the scrub bar — so the UI is a function of the schema, not of parsing free text.
03
People upload video of their own bodies. That drove decisions about retention, access scoping and in-app deletion well before App Store review asked about any of it.
Screens
Still to publish
I have not published usage numbers for this app. When I do, they will be figures I can stand behind rather than rounded-up vanity metrics — so for now the evidence here is the product itself, which you can download and use.
Building something like this?