Camera feed processed entirely on your device using a neural network (face-api.js, trained on FER2013). No images or video leave your device at any point.
Speech is transcribed locally by your browser's built-in speech recognition. The text is analysed entirely on-device using a pattern-matching engine — no data of any kind is sent to any server. Everything stays on your device.
Nothing. All processing — video, audio, speech transcription, and subtext analysis — happens entirely in your browser on your device.