GlassKit gives developers everything needed to build smart-glasses AI apps: a Rokid Glasses starter template, reusable Android patterns for HUD display and input handling, WebRTC streaming helpers, and runnable end-to-end examples that combine camera/mic capture with services like OpenAI Realtime and Overshoot. An agent skill packages all this context so coding agents like Codex, Claude Code, and Cursor can build realistic glasses apps from the first pass.Documentation Index
Fetch the complete documentation index at: https://mintlify.com/RealComputer/GlassKit/llms.txt
Use this file to discover all available pages before exploring further.
Quickstart
Three ways to start: install the agent skill, copy the starter app, or clone a complete example.
Hardware Setup
Get your Rokid Glasses connected via ADB, set up Wi-Fi, and run your first install.
Core Concepts
Understand the four-layer app architecture: glasses, WebRTC, backend, and wearer feedback.
Examples
Explore runnable demos — from IKEA assembly assistants to real-world speedrun timers.
What GlassKit Covers
HUD & Display
480×640 monochrome binocular display patterns, screen controllers, and viewport layout.
Camera & Mic
CameraX back-camera binding, 1024×768 capture, AudioRecord configuration, and WebRTC audio.
WebRTC Streaming
SDP offer/answer flow, peer connection factory, data channels, and ICE setup.
OpenAI Realtime
Low-latency voice-and-vision sessions with backend SDP brokering and sideband control.
Object Detection
RF-DETR inference, normalized detection events, and annotated frame injection.
Voice Commands
Offline Vosk recognition for select, back, next, and previous commands.
How Apps Work
A typical GlassKit app has four pieces that work together:Rokid Glasses (Android)
Captures camera and microphone input, handles touchpad gestures (tap, double-tap, swipe), and renders the monochrome HUD. Built with standard Android APIs plus GlassKit patterns.
WebRTC Transport
Carries live media between the glasses, your backend, and AI services. GlassKit provides the Android client setup, SDP normalization helpers, and Python
aiortc receiver patterns.Backend
Coordinates session setup, workflow state, model calls, tool calls, and app-specific decisions. GlassKit examples use FastAPI (Python) or Node.js depending on the AI service.
GlassKit currently focuses on Rokid Glasses. The long-term goal is a developer platform for building, hosting, and shipping smart-glasses apps across more devices.