Developing for Smart Glasses

An Overview — August 2026

Smart glasses collection — seven AR and AI eyewear devices on a white studio surface

Today we have dozens of smart glasses on the market, each with its own size and feature set. They range from devices with full operating systems to minimal real-time OSes controlled by the phone over Bluetooth.

We can group them into three categories.

The Three Categories

GROUP 1 Full OS — Standalone Apps

The first group runs full Android applications: Snap Spectacles 2024, INMO Air 3, and XReal Aura. All three have very different SDKs. Snap Spectacles uses Lens Studio. XReal Aura uses Android XR. INMO runs stock Android and they both support development with Antigravity or Android Studio.

GROUP 2 Light Android OS — Phone + Optional On-Glasses App

The second group runs a light Android OS and includes the Meta Ray-Ban Display, RayNeo X3 Pro, and Rokid Glasses with Display. They all run on the Qualcomm Snapdragon AR1 processor. These glasses offer two development approaches. The first is to build a phone app that controls the display, buttons, camera, and audio via a custom SDK — typically supporting both iOS and Android. The second is to build a small Android app using the vendor SDK that runs directly on the glasses. In this space, RayNeo X3 Pro is the most open. Rokid requires a custom cable that isn't publicly available, and Meta doesn't offer this option at all.

GROUP 3 Ultra-Light — Phone App Controls Everything

The third group is the ultra-light category: Even Realities G1 and G2, Mentra Live, Vuzix 100, INMO Go 3, and Meta Ray-Ban (non-display model). For these, you build an iOS or Android app using the vendor SDK that runs on the phone, controlling the display, speakers, and capturing button or touchpad input. Vuzix and Even Realities G1 use Mentra OS 2.0. Even Realities G2 and Mentra Live use Mentra OS 3.0 in addition of their own SDK. Meta's SDK is standalone and quite limited in features. All support both iOS and Android.

Smart Glasses Comparison Table

GROUP 1: Full OS — Standalone Apps
Device Processor OS Phone Glasses Display Speaker Camera Edge AI
Snap Spectacles 2024 (5th gen) Dual Snapdragon (models not disclosed) Snap OS 2.0 ❌ ✅ Lens Studio ✅ LCoS waveguide, 46° diag, 37 PPD ✅ ✅ 4 (2 RGB + 2 IR) 🤝 SnapML + OpenAI cloud
INMO Air 3 Qualcomm Snapdragon (model not disclosed), 8 GB / 128 GB Android 14 (IMOS 3.0) ❌ ✅ Play Store (GMS) ✅ Sony micro-OLED 1920×1080/eye, 36°, 600 nits, 120 Hz ✅ ✅ 16 MP, 120° ☁️ ChatGPT
XREAL Aura Snapdragon Reality Elite (puck) + XREAL X1S (glasses) Android XR 🔗 Puck ✅ Android XR ✅ Sony micro-OLED 1920×1200/eye, 70°, 120 Hz ✅ ✅ (count not confirmed) 🤝 Puck NPU + Gemini cloud
GROUP 2: Light Android OS — Phone + Optional On-Glasses App
Device Processor OS Phone Glasses Display Speaker Camera Edge AI
Meta Ray-Ban Display Snapdragon AR1 Gen 1 Meta (proprietary) ✅ Device Access Toolkit ✅ Web Apps (dev preview) ✅ LCoS geometric waveguide 600×600, 20°, 90 Hz, 5,000 nits ✅ ✅ 12 MP, 3× zoom ☁️ Meta AI
RayNeo X3 Pro Snapdragon AR1 Gen 1, 4 GB / 32 GB RayNeo AIOS (Android-based) ✅ ✅ Unity / Android ARDK ✅ Full-color MicroLED waveguide 640×480, 30°, 3,500 nits avg / 6,000 peak ✅ ✅ 12 MP Sony IMX681 + spatial cam ☁️ Gemini
Rokid Glasses Snapdragon AR1 Gen 1 + NXP RT600 Rokid (Android-based) ✅ ✅ SDK ✅ Mono green MicroLED 480×398/eye, 30°, 1,500 nits ✅ ✅ 12 MP Sony IMX681, 109° 🤝 RT600 local voice + cloud
GROUP 3: Minimal — Phone App Controls Everything
Device Processor OS Phone Glasses Display Speaker Camera Edge AI
Even Realities G1 Not disclosed Even OS; MentraOS (3rd-party) ✅ SDK ❌ ✅ Mono green MicroLED waveguide ❌ ❌ ☁️ Phone
Even Realities G2 Ambiq Apollo510b (community teardown, unofficial) Even OS; MentraOS (3rd-party) ✅ SDK ❌ ✅ Mono green MicroLED 640×350, 27.5°, 60 Hz, 1,200 nits ❌ ❌ ☁️ Phone
Vuzix Ultralite Z100 Not disclosed Vuzix Connect; MentraOS (3rd-party) ✅ SDK ❌ ✅ Mono green MicroLED 640×480, 30°, 4:3, right eye ❌ ❌ 🔗 Paired phone
Meta Ray-Ban (Gen 2, no display) Snapdragon AR1 Gen 1 Meta (proprietary) ✅ Device Access Toolkit ❌ ❌ ✅ ✅ 12 MP ☁️ Meta AI
RayNeo iO Not disclosed RayNeo (proprietary) ✅ ❌ ✅ Mono green MicroLED 23.5°, 1,300 nits ❌ ❌ ☁️ RayNeo AI / Gemini / ChatGPT / Claude / DeepSeek
Legend
📱 PhoneApp runs on the phone
👓 GlassesApp runs on the glasses
✅Yes (full support)
❌No
🔗Via companion / puck / tethered
☁️Cloud-based AI
🤝Hybrid (on-device + cloud)

About the author: Philippe Lewicki has been building AR experiences since 2015, working with Microsoft HoloLens, Meta Quest, and now smart glasses across multiple verticals. He consults on smart glasses strategy and private AI backends through AfterNow.

Last updated: August 28, 2026 · View on web

LinkedIn X / Twitter Reddit