Architecture

Monolook is composed of several modules that work together to deliver cross-platform XR experiences from a single codebase.

Main components

Monolook Image

Proprietary computer vision engine compiled to WebAssembly. It detects and tracks images in real time directly in the browser, with no dependency on external services.

• Visual feature detection
• Descriptor matching
• 6DoF pose estimation (full position and orientation in 3D space)
• Simultaneous multi-target
• Binary size: < 3 MB

Monolook World

World Tracking product for placing 3D content on real-world surfaces.

• Android: direct WebXR integration.
• iOS: native bridge through Monolook App/App Clip and ARKit.
• XR headsets: coming soon.
• Unity: coming soon.

License system

World and Image use separate license keys. World validates in monolook.start(). Image validates in tracker.init().

Adapters

Integration layers for the most popular 3D engines:

• ThreeAdapter — Three.js
• BabylonAdapter — Babylon.js
• AFrameAdapter — A-Frame

Each adapter connects the product with the 3D engine, managing camera, scene, and AR content.

General flow

[SDK initialization + license validation]
              ↓
[Load targets (images) or start World Tracking]
              ↓
[Camera active + real-time tracking]
              ↓
[Adapter translates pose → 3D engine]
              ↓
[3D content rendered over camera]