Skip to content

Architecture

Overview

Maverick AI is wirelessly tethered to your smartphone via Bluetooth.

The Maverick AI SDK runs as a part of your mobile application, as simple as that.

There's no need to develop a separate app or adapt to a new development environment. Simply add the Maverick AI SDK to your application, and use it to display your existing app data on the glasses.

The SDK handles all aspects of the interaction with the glasses - from Bluetooth communication to rendering commands and receiving sensor data - and exposes a modern, intuitive API for you to build Maverick AI-enabled applications.

System Overview

┌─────────────────────────────────────────────────────────────┐
│                     Your Mobile App                         │
│                                                             │
│   ┌─────────────────────────────────────────────────────┐   │
│   │                  Maverick AI SDK                     │   │
│   │                                                     │   │
│   │  ┌───────────┐ ┌───────────┐ ┌───────────────────┐  │   │
│   │  │ Glasses   │ │ Screen    │ │ Camera / Mic /    │  │   │
│   │  │ Service   │ │ Service   │ │ Sensors / ET      │  │   │
│   │  └───────────┘ └───────────┘ └───────────────────┘  │   │
│   │  ┌───────────┐ ┌───────────┐ ┌───────────────────┐  │   │
│   │  │ Display   │ │ Resources │ │ OTA / Logger /    │  │   │
│   │  │ Service   │ │ Service   │ │ Preferences       │  │   │
│   │  └───────────┘ └───────────┘ └───────────────────┘  │   │
│   └──────────────────────┬──────────────────────────────┘   │
└──────────────────────────┼──────────────────────────────────┘
                           │ BLE 4.2+ (BLE 5 recommended)
                           │
              ┌────────────▼────────────┐
              │    Maverick AI Glasses   │
              │                         │
              │  ┌───────────────────┐  │
              │  │  Embedded OS      │  │
              │  │  Render Engine    │  │
              │  │  (~67 Hz)         │  │
              │  └───────────────────┘  │
              │  ┌─────┐ ┌─────┐       │
              │  │ IMU │ │ Cam │ ...   │
              │  └─────┘ └─────┘       │
              └─────────────────────────┘

On-Device Rendering

Unlike traditional phone-rendered HUDs, Maverick AI uses an on-device rendering model:

  • Draw objects and their parameters (position, size, color, resources) are sent from the SDK to the glasses once as definitions.
  • The glasses' embedded OS continuously renders frames on-device at ~67 Hz without further phone involvement.
  • The client only sends lightweight incremental updates: object property changes, animator start/stop commands, camera/mic/LOS toggles.

This architecture dramatically reduces Bluetooth bandwidth usage and enables high-quality, high-precision Line of Sight tracking and rendering, since the on-board render loop is decoupled from BLE latency.

Data Flow

Phone (SDK)                              Glasses (Firmware)
    │                                         │
    │── 1. Connect (BLE handshake) ──────────>│
    │                                         │
    │── 2. Object definitions ───────────────>│  Create on-device
    │      (text, shapes, images, resources)  │  scene graph
    │                                         │
    │                                         │── 3. Render at ~67 Hz
    │                                         │      (continuous, autonomous)
    │                                         │
    │── 4. Property deltas ──────────────────>│  Apply changes
    │      (position, color, text, etc.)      │
    │                                         │
    │<─ 5. Events ───────────────────────────-│  Sensor data, LOS
    │      (quaternions, camera, mic, touch)   │  quaternions, camera
    │                                         │  frames, touch events

Key insight: After the initial setup, the phone only sends small deltas when properties change. The glasses handle all rendering independently, making the system resilient to BLE latency and bandwidth constraints.

See Also