OnPar

One phone video of your swing in, a 3D swing analysis and the drills that fix it out.

Role
Founder and only developer. Product, design, the app and the computer vision.
When
Founder · Berkeley and Lund · Nov 2025 – present

The problem

Golf lessons are expensive and launch monitors cost more than most clubs. So most golfers practice alone, film themselves on their phone, and then have no idea what they’re looking at. I wanted the phone video to be enough.

What I built

OnPar turns one swing video into a 3D swing analysis: pose estimation, club tracking and a swing score, then a short list of drills for the thing that matters most. I’m the only developer, so I own everything from the roadmap and the brand to the iOS capture code and the pose backend.

Decisions I’m proud of

One idea per view. I wrote a rule for every screen: one title, one primary action, at most one text link and about a dozen words. If you can’t say what a screen told you after two seconds, it gets cut. The only exception is Measurements, which is dense on purpose.

Don’t show a score you can’t stand behind. One test clip showed “81 Solid” from only 4.5% of the checks. That looks great and means nothing, so below a coverage threshold the app now shows only the metrics it actually measured, without an overall score.

Capture it right instead of fixing it later. iOS hands apps a re-encoded 30 fps file for slow motion, which breaks tempo. So the app records at 240 fps straight to a file, and the backend only trusts tempo when the frame timing checks out.

Where it is now

Getting ready for the App Store and Google Play launch.