Air DJ
A gesture-controlled DJ board that replaces every knob, fader, and hotcue with a hand motion tracked through a webcam.
Date
2026
Role
Computer Vision
Machine Learning
Claude Code
The idea
DJing with physical controllers is wonderful but expensive. Most people have laptops instead, and laptops are cheap. I wanted to collapse the gap by using the one camera everyone already has, training a custom gesture model on my own hands, and routing the classified motions out as MIDI so any DJ software can consume them.
How it works
A MediaPipe hand tracker feeds 21 landmarks per hand into a gesture classifier I trained on data I collected myself. The classifier output maps to virtual MIDI messages routed into VirtualDJ, which handles decks, beatgrids, and the actual audio.
The debug overlay with the hand skeleton exists to show an audience how the system sees them. During an actual set, it is closed. The interface is the gesture, the music, and the room.
Takeaways
Creativity is the real leverage with Claude Code.
The agent is a strong executor but a mediocre inventor. It happily builds whatever you describe, so the value is in deciding what to build. Claude did the typing. The framing was mine.
Bridge existing tools, do not replace them.
Air DJ does not rebuild a DJ engine. It outputs virtual MIDI so VirtualDJ handles decks, beatgrids, and mixing. One small app plugging into a mature one beats reinventing the stack.
Gesture design was the hard part.
The ML was not the bottleneck. The real work was inventing motions that felt obvious on the first try. The crossfader became crossed fingers sliding side to side because the word and the motion match. Every gesture had to carry that kind of immediate meaning or it got cut.
Status
Working prototype. Runs as a standalone macOS app. Next up is hardening the gesture vocabulary and tightening the mappings into VirtualDJ.