Asanafy — Home yoga with real-time AI posture correction

Home yoga practice has a huge blind spot: no one corrects you. Without an expert eye, posture mistakes harden into bad habits that erase the benefits and, in worst cases, cause injuries. Motivation drops because there's no sense of progress, and most people quit within three weeks. The problem isn't lack of videos — it's lack of feedback.

Asanafy is a web-app that turns on the device camera and analyses the pose in real time via pose estimation. When the model detects a misalignment — hips too low, knee off axis, contracted shoulders — it provides an immediate, visual, non-intrusive correction. Gamification turns daily practice into a measurable journey: streaks, milestones, unlockable sequences. Full-web architecture, no downloads, accessible from any device with a camera.

Asanafy is currently in active development. The internal testing phase is validating model accuracy on standard asanas and the UX of real-time feedback. Public roadmap with quarterly milestones and pre-launch community.

Working on computer vision applied to the human body teaches technical humility: what feels obvious to a teacher requires sophisticated models to be captured by a camera. The real discovery was that most users don't want clinical analysis — they want a gentle nudge at the right moment. Feedback design matters as much as model accuracy.