Key Development Takeaway:
Loading…
Upload workout clips or game footage. Review slow-motion mechanics, draw telestrator angle lines, and run AI Vision diagnostics to detect elbow flare, release arc, and foot balance.
Upload a clip, then hit ⚡ Analyze Shooting Form. The video plays through while MediaPipe measures elbow angle, elbow flare, knee dip and release extension on every frame.
Click or drag a clip here — MP4, MOV, WebM. Stays on your device.
🎥 How to film so the analysis works — 60-second read
The two that matter most: film from straight ahead, and get 5–10 shots in one clip. Those alone unlock every measurement.
- Stand in front of the shooter, not beside them. Elbow flare can only be measured front-on. The app checks shoulder width against torso length to work out the angle, and skips flare rather than guess when you're side-on.
- Film 5–10 shots in one take. Set-point consistency — how much the release varies between reps — is the most useful number here, and it needs at least two attempts. One shot gives you "needs 2+ reps".
- Keep the whole body in frame, head to feet, all the way through the follow-through. Joints that leave the frame are skipped, never estimated, so a clip that crops the hands at the top loses the release.
- Get close enough that the player fills the frame. A shooter far away across a gym gives noisy joint positions and shaky angles.
- Start recording before the gather. The app looks back 0.7 seconds from the moment the ball goes up to find the set point, dip and load. Clip the start and it measures the wrong thing.
- One player in frame. The pose tracker follows a single person — a rebounder standing nearby can steal it mid-clip.
- Prop the phone up or use a tripod. Panning and hand-shake move the reference points the angles are measured against.
- iPhone: switch to "Most Compatible". Settings ▸ Camera ▸ Formats, or in Photos use Share ▸ Options ▸ Most Compatible. iPhones record HEVC by default and most browsers can't decode it.
Clips never leave your device — analysis runs in this browser, and only the ~5KB summary is ever saved.
Don't want to tap shots during a live game? Choose the input method that works best for you—speak a quick summary after the game, snap a photo of the paper scorebook, or tap big bleacher buttons!
Speak or type a quick summary of the game. Our AI will automatically parse points, shot locations, free throws, and hand usage into the database!
Notice something during a game or practice? Tap mic or click a quick tag below to feed Qualitative Observations straight into the AI Workout Generator!
- 🗣️ "Misses left hand layups under contact" Factored into AI Drill #2 ✅
- 🗣️ "Doesn't shoot real well when running the floor (transition)" Factored into AI Drill #3 ✅
Snap a photo of the official gym scorebook or paper stat sheet after the game.
Tap to Upload Scorebook Photo
Supports JPEG, PNG (Phone Camera)Huge touch targets designed for tapping from the bleachers while watching live.
When your trainer logs a workout or workout shots in their trainer app, it automatically syncs to your database in real time.
MJ-2027Ranks what to work on by the size of the gap in this athlete's own logged shots. Where a sample is too thin to support a finding, it says so instead of filling the space.