prasanthsasikumar/hermes-glasses — explained in plain English
Analysis updated 2026-05-18
Talk hands free to an AI assistant through Meta Ray-Ban smart glasses with live transcription.
Ask the AI to describe what you are looking at using the glasses' camera.
Connect to a full agentic Hermes Agent on a Mac for tool use and memory across conversations.
| prasanthsasikumar/hermes-glasses | akshaymaloo/filament | crafcat7/peakmon | |
|---|---|---|---|
| Stars | 7 | 7 | 7 |
| Language | Swift | Swift | Swift |
| Setup difficulty | hard | easy | easy |
| Complexity | 4/5 | 2/5 | 3/5 |
| Audience | developer | general | developer |
Figures from each repo's GitHub metadata at analysis time.
Requires iOS 17+, Xcode 16+, Meta Ray-Ban glasses, a Meta App ID and Client Token, and for bridge mode a Mac with Python and the Hermes Agent CLI.
Hermes Glasses is an iPhone app that lets you talk to an AI assistant through Meta Ray-Ban smart glasses, hands free. You can have a spoken conversation with live on-device transcription showing your words as you speak, and you can ask the assistant to describe what you are looking at, using the glasses' camera to take a photo and send it to the AI. The app supports two ways of connecting to an AI. In the simpler Direct mode, the phone talks straight to a provider you choose, such as Claude, OpenAI, Gemini, or a locally running Ollama model, using your own API key, with no extra server needed. In the more advanced Bridge mode, the phone instead connects over a WebSocket to a small Python program running on a Mac, which forwards your questions to a Hermes Agent, a separate tool-using AI system with memory across turns, and streams the spoken reply back to the glasses. On the phone side, the app uses Meta's Wearables Device Access Toolkit to register and connect to the glasses and capture photos, and Apple's built-in speech recognition to transcribe speech entirely on the device, so audio is not sent anywhere just to convert it to text. Replies are spoken back using text-to-speech, either on the phone or through the Mac bridge depending on which mode is active. A built-in test panel lets you check each part of the system, such as the bridge connection, photo capture, and voice queries, separately. To build and run this project you need an iPhone with iOS 17 or newer, Xcode 16 or newer, and a pair of Meta Ray-Ban glasses paired with the Meta AI app. You also need to register for a Meta App ID and Client Token through Meta's developer portal for the glasses to connect. Using Direct mode only requires an API key from your chosen AI provider, while Bridge mode additionally needs a Mac running Python 3.11 or newer with the Hermes Agent command line tool installed.
An iPhone app that lets you talk hands free to an AI through Meta Ray-Ban glasses, with live transcription and camera based visual questions.
Mainly Swift. The stack also includes Swift, SwiftUI, Python.
No license information is provided in the README.
Setup difficulty is rated hard, with roughly 1h+ to a first successful run.
Mainly developer.
This repo across BitVibe Labs
Don't trust strangers blindly. Verify against the repo.