Hand Tracking Reaction Game

Machine learning models trained on millions of hand images can identify hand position, finger orientation, and gesture shape in real time from a standard video stream. This makes it possible to use your hand as a controller with no additional hardware.

Our game runs the hand tracking model on-device using WebAssembly, achieving 30 fps processing without any server round-trip. This latency is low enough to be imperceptible — you will not notice any lag between your hand movement and game response.

Physical engagement increases immersion and performance motivation. Studies in sports psychology show that whole-body involvement in a task increases perceived effort and enjoyment compared to purely digital interaction. This is why arcade cabinets with physical buttons, motion-controlled games, and VR consistently rate higher in engagement than identical games on traditional controllers.

Playing a reaction game with your actual hands creates a more meaningful connection between your physical performance and your score. Improvement feels more real because it involves actual body coordination improvement, not just learned digital dexterity.

The game runs in any modern browser with JavaScript enabled and a working camera. No plugins, no extensions, no downloads. Minimum camera requirement: 720p at 24 fps (standard in almost all laptop webcams and phone cameras manufactured after 2015). Recommended: 1080p for better detection accuracy, though 720p works well in good lighting.