Air Theremin — Play Music with Your Hands Using Your Webcam
Play a theremin without touching anything. Your left hand height controls the pitch — raise it for high notes, lower it for bass. Your right hand distance from your body controls the volume — stretch it out for loud, pull it in to go quiet. No instrument, no install — completely free and private. All processing happens in your browser.
Click here to Start
✋ Left — Pitch
—
Raise hand = higher
🎵 Note
—
C major pentatonic
🤚 Right — Volume
—
Extend arm = louder
Waveform
Scale (C Major Pentatonic)
Prefer clicking instead of waving? Try the Online Piano or Virtual Kalimba.
How to Play the Air Theremin
- Click Enable Sound first — browsers require a tap to unlock audio. You'll hear a test note (A4) and the button turns green.
- Click the camera placeholder above — allow camera access when prompted. Sit ~60–80cm from your camera in a well-lit room.
- Raise your left hand to play higher notes, lower it for bass notes. The note name and frequency are shown in real time.
- Extend your right hand away from your body for full volume, or pull it in to go quiet. Sound only plays when both hands are in frame.
- Choose a waveform to change the sound character — Sine for a classic theremin sound, Sawtooth for a brighter tone.
Pitch Guide
Left Hand Height → Note:
- 🙌 Hand raised high = C6 (1047 Hz)
- 🤚 Mid-high = G5 / A5
- ✋ Mid = C5 / D5 / E5
- 👋 Mid-low = G4 / A4
- ⬇️ Hand low = C3 (131 Hz)
Tips for best results:
- Use a bright, evenly lit room
- Keep both hands visible in frame at the same time
- Move your left hand slowly for smooth pitch glides
- Watch the pitch bar to find your target note
How It Works
The Air Theremin uses a real-time hand tracking model running entirely in your browser via WebAssembly to detect 21 keypoints on each hand at up to 30 frames per second. Your left hand's wrist Y position is mapped continuously to a note in the C major pentatonic scale (16 notes from C3 to C6). Your right hand's wrist X position controls volume — how far your arm extends away from your body. Sound is generated using the Web Audio API OscillatorNode, with frequency changes applied via smooth ramps for lag-free pitch glides. A synthetic reverb impulse response adds warmth. All camera frames are processed locally — nothing leaves your device.
The scale is locked to C major pentatonic (C, D, E, G, A) rather than full chromatic, which means every position sounds musical — no accidental "wrong" notes.
Frequently Asked Questions
Does this send my camera feed to a server?
No. All hand tracking and audio synthesis runs entirely in your browser using JavaScript and WebAssembly. Your camera feed never leaves your device. No data is stored or transmitted.
Why is there no sound?
Sound only plays when both hands are detected in frame simultaneously. Also make sure you've clicked Enable Sound first — browsers block audio until you interact with the page. The button turns green when audio is active.
Why does the pitch jump instead of glide smoothly?
The scale is quantised to 16 notes (pentatonic). Move your hand more slowly for smoother transitions between adjacent notes. The frequency transition itself uses a 20ms smoothing ramp to avoid clicks.
What does the waveform selector do?
It changes the timbre (tone colour) of the sound. Sine is the purest, closest to a real theremin. Triangle is softer and flute-like. Sawtooth is bright and buzzy. Square has a hollow, woody character.
Can I use this on mobile?
Yes — the tool works on modern mobile browsers with a front-facing camera. Use Chrome or Safari on a recent device. Hand tracking is CPU-intensive on mobile, so performance may vary. Landscape orientation gives more hand movement room.
What scale is used?
C major pentatonic — the 5 notes C, D, E, G, and A, spread across three octaves (C3 to C6). This scale was chosen because every note combination sounds harmonious, making it easy to play melodies without any music theory knowledge.