Notch icon

Notch

Score any match card from a photo with on-device ML

Download on the App Store
Read the Story
My Phone Replaced a Brass Plug I wanted to cook venison from scratch, which meant learning to shoot, which meant keeping track of my progress, which meant porting a 2012 OpenCV paper and training a state-of-the-art computer vision model, which meant the dinner took a bit longer than expected.
→
Notch scoring a photographed match card: 89/100, with breathing, sights and fatigue flags A live range session on iPhone: 437 of 500 points across five targets Shot placement on one target, with group size, spread and a score breakdown by ring Statistics ranking breathing as the most common fault, with a breathing pattern chart Season stats: score progress over time, best session, tightest group and score distribution A live session on the Lock Screen showing 24 shots at 92.5%

Photograph the target. Notch scores the card, spots the drift in your group, and tells you which fundamental to work on next — so the drive home isn't the first time you think about what just happened on the firing line.

Core ML or Vision framework work? I consult on on-device ML integrations.

Common questions

What target types does Notch support?
NSRA .22 cards are the primary format. ISSF and air-rifle targets work too, with more disciplines added over time.
Does Notch need internet?
No. CoreML runs the scoring on-device. No photos leave your phone.
Can Notch tell me what I'm doing wrong?
It flags common group-shape patterns — a tight cluster low and right usually means trigger pull; a vertical string through centre is breathing through the shot. Cumulative heat maps show the patterns across sessions.
Can I compare across different rifles or range days?
Yes. Every scored card is saved with rifle and date. Trend analysis across sessions shows per-rifle performance and whether this week's groups are tighter than last week's.
Terms & Privacy