RoundKeep detects shot groups from a photo of your target. By default you get the point-of-impact map plus consistency and horizontal / vertical spread; calibrate the scale and you also get group size in inches or MOA.
Detection runs on-device, on a small ML model trained on real target photos. Nothing is uploaded, nothing is collected — the photos and the analysis stay on your phone.
How it works
- Take a photo of the target. Stand reasonably square to it, get the whole thing in frame.
- Attach the photo to a practice session record.
- The app detects impact points automatically.
- Confirm distance so group sizes can be reported in inches or MOA.
- Calibrate scale — drag the reference circle so it tightly hugs one of the bullet holes. That sets the size reference for every measurement on the photo.
When detection misses
It will, sometimes — paper has shadows, lighting varies, and the model is still being trained on more targets. Tap Edit Detections and:
- Tap empty space to add a hole.
- Tap a hole to select it, then delete.
- Assign holes to specific groups manually.
- Pinch to zoom for precision.
Group sizes recalculate live as you correct. The user is always in the loop.
Where it fits
Each analysis attaches to a target photo on a practice entry. Your group sizes live alongside the firearm and ammo combo that produced them — so you can scan back and see whether your 124gr Federal HST groups tighter than 115gr American Eagle out of a particular pistol.
