Open source smart glasses: AI glasses and Meta Ray-Ban apps
Last updated
Eight open-source smart glasses projects, from DIY glasses you build yourself to open apps and SDKs for Meta Ray-Ban and Rokid glasses. Each is described in one line quoted from its creator's repository at a pinned commit, with its license and two or three facts linked to the exact README lines. Two of the most-starred builds, OpenGlass and Open Source Smart Glasses, are no longer maintained in their original repositories; their entries say so.
1. OpenGlass
“Turn any glasses into hackable smart glasses” (source)
- Hardware: a Seeed Studio XIAO ESP32 S3 Sense, a 3.7 V 250 mAh battery and a 3D-printed glasses mount. (source)
- AI: API keys for Groq and OpenAI, or a self-hosted Ollama server running the moondream vision model. (source)
- Status: the README says the project moved to the Omi repository and this repo isn't supported anymore. (source)
RepositoryLicense: MITCredit: Based Hardware
2. MentraOS
“MentraOS is how developers and businesses build smart glasses apps.” (source)
- What it does: handles pairing, connection, data streaming, hardware access and cross-device compatibility for glasses apps. (source)
- Supported glasses listed in the README include Mentra Live, Even Realities G2, Vuzix Z100 and Even Realities G1. (source)
- License scope: the README says every component is open source under the Apache License 2.0. (source)
RepositoryLicense: Apache-2.0
3. Open Source Smart Glasses
“This repo holds mechanical, electrical, and software files for the open source smartglasses developed by Team Open Smart Glasses.” (source)
- Hardware: smart glasses with a display, microphones, a wireless phone connection, prescription lenses and an LED indicator. (source)
- Status: a 2024 update says the team moved on to build AugmentOS, an OS for glasses such as the Even Realities G1 and Vuzix Z100. (source)
- You build your own: the README says there is no way to purchase a pair. (source)
Source-listed parts
As an Amazon Associate HardwareLab earns from qualifying purchases.
- Find ESP32-PICO-D4 on Amazon (source)
U2 in the v0.5.5 PCB BOM; this is a chip, not a complete glasses kit.
RepositoryLicense: MITCredit: Team Open Smart Glasses
4. OpenSource AI Glasses
“An open-source Linux-based smart glasses platform for medical, industrial, educational, and consumer applications” (source)
- Hardware: any RV1106B-based development board can be the starting point for the open system and software stack. (source)
- Camera: the glasses can stream their camera feed over RTSP for the mobile client and for cloud-side video AI recognition. (source)
- Status: the README calls it early-stage, at 45% documentation completeness. (source)
Source-listed parts
As an Amazon Associate HardwareLab earns from qualifying purchases.
- Find RV1106B development board on Amazon (source)
The README says peripheral adaptation is needed for your own board.
RepositoryLicense: Apache-2.0
Into smart glasses? Get the Thursday email
Five hardware builds every Thursday. Email only, no tracking cookies, unsubscribe in one click. We never share your address.
Built something? Submit a build
5. GlassKit
“GlassKit is an open-source toolkit with an agent skill, development guides, runnable examples, and vision evals for apps that understand what wearers see and hear, then guide them in real time.” (source)
- Glasses: device integrations and examples currently focus on Rokid Glasses. (source)
- Building blocks: camera and mic capture, WebRTC streaming, a monochrome on-lens HUD, OpenAI Realtime and object detection. (source)
RepositoryLicense: MIT
6. OpenVision
“The open-source iOS app connecting Meta Ray-Ban smart glasses to AI assistants — cloud or fully on-device.” (source)
- AI: seven backends, cloud or fully on-device. (source)
- On-device models via Apple MLX: Qwen 2.5 (0.5B/3B), Gemma 2 2B, Gemma 4 E2B and SmolVLM2 2.2B. (source)
- Live voice and vision: Gemini Live with native audio streaming. (source)
RepositoryLicense: MIT
7. Rokid AI Assistant
“AI-powered voice and vision assistant for Rokid AR glasses.” (source)
- Camera: photo capture from the glasses camera with AI image analysis. (source)
- AI: 15 providers, including Gemini, OpenAI and Anthropic. (source)
- A phone is required: standalone glasses-only operation is out of scope. (source)
RepositoryLicense: Apache-2.0
8. xg.glass SDK
“xg.glass is a fast and easy-to-use library for smart glasses application development.” (source)
- One API for four interfaces: camera video in, microphone audio in, display out and audio out. (source)
- Glasses in its support table include Rokid Glasses, Meta Wearables (video marked as a phase 2 validation target), Brilliant Labs Frame and RayNeo X2. (source)
RepositoryLicense: Apache-2.0
Ray-Ban Meta chess coach: AI best-move guidance through smart glasses →