GlassFlow
Live transcription and AI assistant on Ray-Bans.
By Intent-Lab
- Est. cost
- Cost unknown
- Posted
- GitHub stars
- 91 (GitHub)
Overview
Real-time transcription for Meta Ray-Ban smart glasses: live speech-to-text with speaker labels from the glasses microphone, plus an optional Gemini Live assistant.
How it works
Audio from the glasses (or the phone mic) streams to Deepgram Nova-3 over a WebSocket, which returns partial and final results and labels speakers S1, S2 and so on. Gemini Live adds optional voice and vision AI, and WebRTC shares your view to a browser. It runs on iOS and Android.
Hardware
- Meta Ray-Ban glasses
This page is a short summary written by HardwareLab. Build steps, code and photos belong to the creator; follow “View original” for the full project.
More in Smart Glasses
See all 31Hermes Glasses
Voice + vision assistant for Ray-Bans and Realtek-based glasses.
- Meta Ray-Ban
- AiSee/Realtek glasses
AdvancedCost unknown22Visionbridge AI Glasses
ESP32 glasses firmware with voice, camera, display and MCP.
- ESP32
- camera
- display
- +1 more
AdvancedCost unknown5Visionclaw
Modular glasses: stock frame plus 3D-printed AI temples.
- XIAO ESP32-S3
- bone-conduction audio
- 3D-printed temples
IntermediateCost unknown2