lens-api

Bind a TapEvent with this.createEvent('TapEvent') and read eventData.getTapPosition(), which returns normalized [0-1] coordinates. For drag or multi-touch, use InteractionComponent's onTouchStart, onTouchMove, and onTouchEnd callbacks instead.

Call camera.screenSpaceToWorldSpace(tapPos, depth) with the normalized tap position and a depth measured from the camera's near plane. The tap position and this method share the same coordinate system, so no manual projection math is needed.

Materials are shared assets, so modifying mainMaterial in place recolors every object using it. Clone the material first, call clearMaterials() and addMaterial(clone) on the RenderMeshVisual, then modify the clone.

The Lens API (StudioLib) runs inside the AR experience, while the Editor API (editor.d.ts) drives the Studio editor. They use different accessors — for example getTransform().getLocalPosition() versus localTransform.position — and mixing them compiles but fails silently at runtime.

Cross-object @input references and getComponent calls on other objects are unsafe during onAwake because initialization order is not guaranteed. Defer that work to OnStartEvent, which fires after all components have finished onAwake.

Call global.scene.createSceneObject('name') on the ScriptScene, optionally set a parent, then add components with createComponent('Component.RenderMeshVisual'). SceneObject itself has no creation method — global.scene is the only runtime entry point.