Apple's InputTargetComponent documentation shows this code for an entity that should receive input without participating in physics: // Create a collision component with an empty group and mask. var collision = CollisionComponent(shapes: [.generateSphere(radius: 0.1)]) collision.filter = CollisionFilter(group: [], mask: []) myEntity.components.set(collision) In my RealityKit scene, myEntity also has an enabled InputTargetComponent and a GestureComponent with a TapGesture callback. I use .trigger collision mode. With the empty group and mask, the callback runs on visionOS but does not run on macOS or iOS. Changing the filter to CollisionFilter.default makes the callback run on macOS and iOS. The entity and gesture setup are otherwise the same. Should the documented empty filter still allow input hit testing on macOS and iOS? Is this a RealityKit bug, or does input targeting on those platforms require a different collision filter or additional setup?

I filed Feedback Assistant report FB24967734 for this issue. When two USD scenes contain explicit polygon meshes at the same prim path and are compiled together with xcrun realitytool, one scene's mesh loads with zero bounds. Renaming the mesh prims so their paths are unique makes both meshes load correctly. Reproduce Download and expand the source ZIP from https://github.com/mesqueeb/RealitykitUsdMeshDisappearanceOnMatchingPaths/archive/refs/heads/main.zip, then run these commands from the extracted directory: xcrun realitytool compile --output-reality Fixtures/Compiled/inline-same-path.bundle/Contents/Resources/RealityKitContent.reality --platform macosx --deployment-target 27.0 Fixtures/Source/inline-same-path.rkassets xcrun realitytool compile --output-reality Fixtures/Compiled/inline-unique-path.bundle/Contents/Resources/RealityKitContent.reality --platform macosx --deployment-target 27.0 Fixtures/Source/inline-unique-path.rkassets swift run The same-path case has a mesh at /Root/MeshCube in both scenes. The control changes only the mesh paths to /Root/MeshCube_A and /Root/MeshCube_B. Current output inline-same-path: SceneA.usda / MeshCube: bounds=SIMD3<Float>(0.19999999, 0.2, 0.2) SceneB.usda / MeshCube: bounds=SIMD3<Float>(0.0, 0.0, 0.0) inline-unique-path: SceneA.usda / MeshCube_A: bounds=SIMD3<Float>(0.19999999, 0.2, 0.2) SceneB.usda / MeshCube_B: bounds=SIMD3<Float>(0.19999999, 0.2, 0.2) Expected output Both meshes in the matching-path case should have nonzero bounds: inline-same-path: SceneA.usda / MeshCube: bounds=SIMD3<Float>(0.19999999, 0.2, 0.2) SceneB.usda / MeshCube: bounds=SIMD3<Float>(0.19999999, 0.2, 0.2) inline-unique-path: SceneA.usda / MeshCube_A: bounds=SIMD3<Float>(0.19999999, 0.2, 0.2) SceneB.usda / MeshCube_B: bounds=SIMD3<Float>(0.19999999, 0.2, 0.2) Reproduced on macOS 27.0 (26A428), Xcode 27.0 (27A266a). The repo includes the USDA inputs, compiled RealityKitContent assets, and the Swift code that prints these bounds. Source and downloadable ZIP: https://github.com/mesqueeb/RealitykitUsdMeshDisappearanceOnMatchingPaths

On visionOS 26.6 and 27 latest beta: PresentationComponent in a volumetric WindowGroup, when looking at the volume from a specific lower/back region create two bugs: buttons inside the presentation have their gaze obstructed by hand presence, no button shows any hover highlight hand movement in that specific area auto dismisses the presentation The focused public repro uses a thin ModelEntity with InputTargetComponent and ordinary SwiftUI buttons. Its README includes two clearer videos: https://github.com/mesqueeb/PresentationAutoCloseInBackOfVolume 1. Gaze obstructed on buttons when hand is at specific location (see "Test Action" has no hover highlight, even though i look at it) 2. Hover state restored on buttons when hand is gone (see "Test Action" has hover highlight) 3. Presentation auto closed when hand comes back The weirdest part: This only happens at a specific location, when walking to the back of the volume, then turning around and interacting with the volume from the lower right back side. Any workaround or fix guidance would be appreciated. Feedback Assistant: FB24569418.

Setting massProperties on a USDZ-loaded entity corrupts its transform (position/scale → NaN) Category: RealityKit / visionOS Environment: visionOS 26.x Summary: Modifying PhysicsBodyComponent.massProperties at runtime on an entity loaded from a .usdz file silently corrupts the entity's transform. The translation and scale components become NaN, while rotation remains valid. The corruption occurs during the next RealityKit scene update cycle (e.g., during an await suspension on the main actor). Steps to Reproduce: Author a .usda file with a RigidBody component including valid m_mass and m_inertia values Load the entity at runtime via Entity(named: "MyEntity.usdz", in: bundle) Read the existing PhysicsBodyComponent, modify massProperties, and set it back: if var physics = entity.components[PhysicsBodyComponent.self] { physics.massProperties = .init( mass: 1.944, inertia: SIMD3<Float>(0.013, 0.026, 0.016), centerOfMass: (position: .zero, orientation: .init()) ) entity.components.set(physics) } Add the entity to the scene Perform any await call (e.g., TextureResource(contentsOf:)) that yields to the main actor, allowing a RealityKit scene update to run Expected: The entity retains its position and scale with the updated mass/inertia values. Actual: The entity's transform becomes corrupted: Transform( scale: SIMD3<Float>(nan, nan, nan), rotation: simd_quatf(real: 1.0, imag: SIMD3<Float>(0.0, 0.0, 0.0)), // ← fine translation: SIMD3<Float>(-nan, -nan, -nan) ) Additional findings from investigation: The entity's physics mode does not matter — corruption occurs even when the entity is .kinematic Creating a new PhysicsBodyComponent(...) and replacing the existing one also triggers the bug Modifying other properties on the same component (linearDamping, angularDamping, material, mode) does not cause corruption — only massProperties triggers it The parent entity's transform remains valid The computed mass/inertia values themselves are valid (finite, positive) The corruption is silent — no error, no warning, no crash from RealityKit itself Workaround: Author mass and inertia values directly in the .usda file and do not modify massProperties at runtime: def RealityKitStruct "massFrame" { float3 m_inertia = (0.02, 0.02, 0.038) float m_mass = 2.5 def RealityKitStruct "m_pose" { } } Other PhysicsBodyComponent properties (damping, material, mode) can safely be modified at runtime.

Filed as FB23362414, with a minimal sample: https://github.com/mesqueeb/Swiftui-Gesture-Detection-Failures While a SwiftUI DragGesture is held (trackpad click-drag), a two-finger magnify (trackpad pinch) permanently cancels it: onChanged stops firing the instant the magnify is recognized. onEnded is never delivered — the gesture is torn down with no terminal event. Continued motion of the same, still-pressed finger after the pinch is not re-detected. The drag only recovers after a full release and re-press. Throughout, the AppKit NSEvent stream keeps delivering .leftMouseDragged (and a clean .leftMouseUp on release), so the OS is still tracking the drag — it's SwiftUI's gesture arbitration that discards it. No gesture composition avoids this: .simultaneousGesture, .highPriorityGesture, varying gesture order, and .exclusively(before:) in both directions were all tried; none delivers a terminal onEnded or resumes the drag after the pinch. Steps to reproduce (full sample in the repo — ./build.sh run): Press-hold and drag a shape with the trackpad (do not release). Without lifting the drag finger, perform a two-finger pinch. End the pinch and keep moving the same finger. Expected: onChanged continues for the still-pressed finger; onEnded fires when it's lifted. Actual: onChanged stops at the pinch and never resumes; onEnded never fires. Workaround: driving the gestures off AppKit NSEvent instead of SwiftUI works correctly — the sample has a toggle to switch between the two so you can compare side by side. Tested on macOS 26.3.1 (25D771280a) and macOS 27.0 Beta (26A5353q), Apple Silicon, built-in trackpad. Has anyone else run into this, or found a SwiftUI gesture composition that survives the pinch?

When running parallel Swift Testing tests that each create their own in-memory ModelContainer, the first xcodebuild test run succeeds but all subsequent runs fail with DefaultStore save failed validation errors — unless DerivedData is deleted between runs. Setup Each test suite creates an isolated in-memory container with a unique name: ModelConfiguration(UUID().uuidString, schema: schema, isStoredInMemoryOnly: true) The @Model class has non-optional properties (e.g. var v: String, var entityDicMoveNr: Int) that are always set in init. Symptoms on second run SwiftData.DefaultStore save failed with error: Error Domain=NSCocoaErrorDomain Code=1560 "Multiple validation errors occurred." NSValidationErrorKey=v, NSValidationErrorValue=null NSValidationErrorKey=entityDicMoveNr, NSValidationErrorValue=null NSValidationErrorKey=historyBackup, NSValidationErrorValue=null The errors appear before test code executes — during SwiftData's internal DefaultStore auto-save. Properties that are always initialized in Swift show as nil in the Core Data layer. What we verified Each run uses a fresh xctest process (different PIDs) Container creation succeeds (no errors in our ModelContainer init) The @Model init always sets all fields — confirmed with debug prints Deleting DerivedData before each run fixes it (but forces a full rebuild) What didn't help Using ModelContext(container) instead of container.mainContext Setting autosaveEnabled = false on both mainContext and custom contexts Using /dev/null URL instead of isStoredInMemoryOnly: true build-for-testing + test-without-building Clearing only Logs/Test or TestResults subdirectories Why not run via swift test? We can't fall back to swift test because our project uses RealityKit custom components, which silently fail to decode in CLI swift test. So xcodebuild test is our only option, making the DerivedData workaround especially costly. Workaround Delete DerivedData before each xcodebuild test invocation. Environment macOS 15.5, Xcode 16.4, Swift Testing with @Suite @MainActor parallel test suites. Question Is there a way to fully isolate in-memory ModelContainer instances so that no metadata leaks between xcodebuild test runs without requiring a full DerivedData clean?

swift test (SwiftPM CLI) fails to decode RealityKit custom components from USD files, even though entity.components.has(MyComponent.self) returns true. Typed access via entity.components[MyComponent.self] returns nil. This forces projects that use RealityKit custom components to use xcodebuild test exclusively. Minimal repro: github.com/mesqueeb/swiftpm-realitykit-custom-component-repro Repro steps git clone https://github.com/mesqueeb/swiftpm-realitykit-custom-component-repro cd swiftpm-realitykit-custom-component-repro swift test --filter componentsPresentButNotDecodableInSwiftTest Observed ✅ entity.components.has(ReproComponent.self) returns true ❌ entity.components[ReproComponent.self] returns nil Expected If has(...) returns true for a registered custom component, typed lookup should decode and return non-nil. Notes Running the same test via xcodebuild test works correctly The component is properly registered and the USDA file correctly references it This affects any project that relies on custom RealityKit components in tests — there is no swift test workaround Feedback ID: FB22099519 Environment: macOS 15.5, Xcode 16.4, Swift 6.1

So with the new ManipulationComponent, we can choose "stay" and then if you drag it out of your volume, once you let go it will instantly disappear. We can "animate" it back to inside the volume, eg.: content.subscribe(to: ManipulationEvents.WillRelease.self) { event in Entity.animate( .easeInOut(duration: 1), body: { event.entity.position = [0, 0.2, 0] }, completion: {} ) }, Howeve,r for the duration that it travels outside of the volume it's invisible the whole time. In this apple video, it seems to be visible when dragging and when letting go, but perhaps that's not a volume they're dragging it out of? https://youtu.be/VtenPKrvPOU?si=y1zoZOs2IMyDzOm6&t=1748 Does anyone know how to keep the entity visible even when after letting the entity go while you animate it back towards inside of your volume?

So if I drag an entity in RealityView I have to disable the PhysicsBodyComponent to make sure nothing fights dragging the entity around. This makes sense. When I finish a drag, this closure gets executed: .gesture( DragGesture() .targetedToAnyEntity() .onChanged { e in // ... } .onEnded { e in let velocity: CGSize = e.gestureValue.velocity } If I now re-add PhysicsBodyComponent to the component I just dragged, and I make it mode: .dynamic it will loose all velocity and drop straight down through gravity. Instead the solution is to apply mode: .kinematic and also apply a PhysicsMotionComponent component to the entity. This should retain velocity after letting go of the object. However, I need to instatiate it with PhysicsMotionComponent(linearVelocity: SIMD3<Float>, angularVelocity: SIMD3<Float>). How can I calculate the linearVelocity and angularVelocity when the e.gestureValue.velocity I get is just a CGSize? Is there another prop of gestureValue I should be looking at?

Can anyone provide or point me to example code to fade in / out spotlights over 1 second? Did not find anything on this topic in the docs: https://developer.apple.com/documentation/realitykit/spotlight

What’s the difference between an action and an animation eg.: FromToByAnimation vs FromToByAction. The documentation on them is pretty similar and I'm not understanding the differences exactly... : S FromToByAnimation → https://developer.apple.com/documentation/realitykit/fromtobyanimation?changes=__2_2 FromToByAction → https://developer.apple.com/documentation/realitykit/fromtobyaction?changes=__2_2 As developer, when should we reach out to use an animation vs action ? 🤔

What's the difference between project dependencies and target dependencies? Project dependencies can be added here: Target dependencies can be added here: in the case of external dependencies (via URL), do we need to add it to both project and target, or is one of those enough? in the case of local dependencies (via the workspace), do we need to add it to both project and target, or is one of those enough?

I have an immersive environment with a skybox which uses an png image inside a sphere. I added an IBL, but I am not sure what the best format / prep method is for the IBL image. I have tried several different images for my IBL, and all are very different vibes from what I have in Blender. My question is how can I create an IBL that's closest to Blender's Cycles rendering engine? However, it's a rather difficult to answer question, so I want to ask some smaller questions first. Does IBL need to be BW or will colour work? From my tests: colour works just as well. But why does Apple only show use of BW ones? Should we use BW? What is the best file format for IBL? Any pros/cons? Or should we just test out each format and check visually. From my tests: PNG, OpenEXR (.exr), Radiance HDR (.hdr) all work. But which format is recommended? Will IBL on visionOS create shadows for us? In Blender an HDRI gives shadows. From my tests: No, IBL does not provide shadows on your loaded environment/meshes. Is "shadow baking" the only solution for the time being? Looking at a scene in Blender which uses HDRI as global lighting, how can we best "prep" the IBL image that will give the closest light similar to Blender's Cycles rendering engine? From my tests: I tried (as shown below) A) make a render of just the Blender HDRI (without meshes) via 360-degree panoramic camera. →　Usage as IBL makes everything too bright. B) make a render of the entire Blender scene via 360-degree panoramic camera. →　Usage as IBL makes everything washed out and yellowish. C) Use the Sunlight.png from the sample project. →　With this IBL the scene is too dark. D) Use the SystemIBL.exr from another sample project. →　With this IBL the scene looks very flat and not realistic at all. Here I show each IBL I described above 1~4 and sample screenshots from the simulator: A) B) C) D) The atmosphere I'm aiming for as per Blender's Cycles rendering engine: Can anyone help me with my questions 1 ~ 4 above. It would give me some insight in how to create immersive environments with realistic lighting & shadows. : ) Much appreciated! — Luca

The error I get with visionOS simulator: cannot migrate AudioUnit assets for current process code: guard let resource = try? AudioFileGroupResource.load( named: "/Root/AudioGroupDropStone", from: "Scene.usda", in: realityKitContentBundle ) Any ideas how to debug this? The audio files seem to work fine in Reality Composer Pro.

SharePlay & Group Activities I was able to implement entity position synchronisation via SharePlay (Group Activities) in my visionOS app by following the tutorials on SharePlay in the "Draw Together" app from these WWDC sessions: https://developer.apple.com/wwdc21/10187 https://developer.apple.com/wwdc22/10140 While referencing the sample code at: https://developer.apple.com/documentation/groupactivities/drawing_content_in_a_group_session MultipeerConnectivityService However, it seems that RealityKit has something called MultipeerConnectivityService for Entity position synchronisation and it seems to be a pretty robust solution that will sync not only positions but also other things like Codable components. 🤔 See docs at: https://developer.apple.com/documentation/realitykit/multipeerconnectivityservice Call for help Can anyone share example code that implements MultipeerConnectivityService ? I wonder if this is the recommended approach by Apple? I must say, writing custom messages to sync the Entity positions via Group Activities was very hard 😅 I was just thinking what I should do for all the entity components now...