What is Edge Processing?
Edge Processing means all data processing happens right on your device — in the browser, with nothing sent to a server.
How it works in Red Rabbit
- . **Camera** → Video is processed by ARKit TrueDepth right on your device
- . **Microphone** → Audio is analyzed with the Web Audio API in your browser
- . **Neural network** → faceExpressionNet runs in your browser
- . **Result** → ONLY a number is sent to the server (for example: score=67%)
What does this mean for you?
- 🔒 Your video NEVER leaves your device
- 🔒 Your voice is NEVER recorded on a server
- 🔒 Even if our server were hacked — there is no video/audio on it
- 🔒 No one can view your recordings — they simply do not exist
Technical implementation
- ARKit TrueDepth: 1,220 3D facial points + 52 blendshapes → 60 FPS → locally
- Web Audio API: RMS amplitude + frequency analysis → locally
- faceExpressionNet: 7 emotions → 2 FPS → locally
- What gets transmitted: a numeric array of ~20 values per question
This is fundamentally different from competitors, who often send video to their servers for analysis.