Badminton AI: Swing Coach

About

Badminton AI turns a video of your forehand clear into focused, actionable form feedback. Choose a video of one complete swing. On your iPhone, the app detects body landmarks, locates preparation, acceleration, and estimated-contact key frames, measures elbow extension and contact-point height, and compares them with rule-based reference ranges. You can review: • Three key frames with body-pose overlays • Your motion aligned with a rule-based reference pose • Quantitative metrics such as elbow extension • A clear link between a motion difference, its likely effect, and what to change • One focused drill for your next practice • Analysis history stored on your device Privacy by design: Video selection, pose detection, and Core ML inference happen on your device. Source videos are not uploaded to a server. When history is enabled, the app stores only three compressed key frames and structured report data, which you can delete at any time. This version focuses on the forehand clear. The bundled model uses shuttlecock detection only as a localization aid; racket-head detection is not available. The reported contact moment is a visual estimate, not confirmation of physical contact. The reference pose is rule-based and is not professional-player footage. Results are for badminton practice only. They are not medical, rehabilitation, or injury-prevention advice. Train in a safe space and adjust activity to your condition.