spatial-gestures-interaction

Attach SwiftUI gestures like SpatialTapGesture or DragGesture to a RealityView and chain targetedToAnyEntity or targetedToEntity. The entity must have an InputTargetComponent and a collision shape, typically via generateCollisionShapes, or the gesture never fires.

The most common cause is a missing InputTargetComponent or missing collision shape on the entity. Set entity.components.set(InputTargetComponent()) and call generateCollisionShapes(recursive: true); without both, gestures silently do nothing.

Capture the entity's starting position in @State when the drag begins, convert the gesture's translation3D into the parent coordinate space with value.convert, and apply the delta to the start position. Reset the stored state in onEnded to avoid jumps.

No, raw gaze is private by design on visionOS. The system resolves the looked-at target and applies hover effects itself; your code only learns what was interacted with through gestures, never gaze direction or position.

Gestures like pinch and tap arrive as input events targeted at entities and need no special permissions. Raw hand tracking with joint positions requires ARKit's HandTrackingProvider, a Full Space, and user authorization, and should only be used when joint data is genuinely needed.