Project
- Stars
- Started
- Issues
- License
Run a camera-processing and voice prototype
FirstSight sends the glasses' camera view through Meta's Wearables Device Access Toolkit to an Android app and a backend you run. Its implemented Fast Whisper mode combines local speech transcription with a configurable Gemini text model, and the backend includes experimental camera-based heart-rate processing.
Build the Android prototype, retain Meta AI pairing and Developer Mode, and configure the backend and provider key. The advertised facial-droop weights are not published, while the separate standalone Gemini Live mode depends on a model with an announced earliest shutdown date. Exact glasses generations and clinical validation are unverified.
How did it go? Tips, problems, questions…
Run a camera-processing and voice prototype
FirstSight sends the glasses' camera view through Meta's Wearables Device Access Toolkit to an Android app and a backend you run. Its implemented Fast Whisper mode combines local speech transcription with a configurable Gemini text model, and the backend includes experimental camera-based heart-rate processing.
Build the Android prototype, retain Meta AI pairing and Developer Mode, and configure the backend and provider key. The advertised facial-droop weights are not published, while the separate standalone Gemini Live mode depends on a model with an announced earliest shutdown date. Exact glasses generations and clinical validation are unverified.
Safi Shamsi and 1 other
2,302 followers · joined Jun 2025
Feedback
How did it go? Tips, problems, questions…