SprinSole
foot 3D scan for custom insoles
I'm the founder. Some context on why this exists and how it actually works, since "AI custom insoles" probably reads like a gimmick.
I'm a computer-vision engineer. I worked at Bodylabs on 3D human body reconstruction (acquired by Amazon in 2017), then at Amazon on fashion projects. I have flat feet and spent years cycling through insoles — drugstore inserts never did much, and genuinely custom orthotics from a clinic ran into the hundreds of dollars (often $300–500+), because each pair involves manual casting and design. The whole goal of this project was to make truly custom insoles affordable for normal people.
The bet is that the expensive parts — capturing foot geometry and designing the support — can be automated, and that's where the cost savings come from. The app captures a ~1mm 3D scan of each foot with the iPhone's TrueDepth sensor: https://apps.apple.com/ph/app/sprinsole/id6757342866 (Or Android uses photogrammetry via Polycam). That scan feeds a generative model that designs the insole geometry per foot — arch height, heel cup — which is then 3D printed and shipped from California, usually within a week. Automating those steps is what brings a genuinely custom pair down to about $75, instead of the several hundred a clinic charges (the storefront is sprinsole.com if you want to see the finished product).
The hard parts, honestly:
Getting a clean scan from a consumer phone. People move mid-scan and you get a doubled surface, so a lot of the app is just detecting and rejecting bad scans before they reach manufacturing.
Making the generated geometry actually manufacturable, not just plausible on screen.
Feet are asymmetric — most people's arches differ left to right — so the two insoles in a pair usually aren't identical, which surprises people.
Limitations I want to be upfront about: this is not a medical device. If you have severe over-pronation or diabetic foot issues, see a podiatrist, not me.
The app is free — you can scan your feet and see your model without buying anything, and that's genuinely the part I'd most like feedback on: where the scan flow breaks, and what would make you trust (or not trust) a custom insole ordered from your phone. If you do want to try a pair, HN folks can use code HACKERNEWS for 10% off — a thanks for taking the time to look.
1 Comment
I think the biggest challenge isn't convincing people the scan works—it's convincing them that a phone can replace what they've always associated with a clinic visit. The engineering gets you a great product, but the adoption hinges on crossing that trust gap. I'd spend as much time validating what creates confidence as improving the scan itself, because that's probably the real bottleneck.