Hi everyone, I’m encountering a memory overflow issue in my visionOS app and I’d like to confirm if this is expected behavior or if I’m missing something in cleanup. App Context The app showcases apartments in real scale using AR. Apartments are heavy USDZ models (hundreds of thousands of triangles, high-resolution textures). Users can walk inside the apartments, and performance is good even close to hardware limits. Flow The app starts in a full immersive space (RealityView) for selecting the apartment. When an apartment is selected, a new ImmersiveSpace opens and the apartment scene loads. The scene includes multiple USDZ models, EnvironmentResources, and dynamic textures for skyboxes. When the user dismisses the experience, we attempt cleanup: Nulling out all entity references. Removing ModelComponents. Clearing cached textures and skyboxes. Forcing dictionaries/collections to empty. Despite this cleanup, memory usage remains very high. Problem After dismissing the ImmersiveSpace, memory does not return to baseline. Check the attached screenshot of the profiling made using Instruments: Initial state: ~30MB (main menu). After loading models sequentially: ~3.3GB. Skybox textures bring it near ~4GB. After dismissing the experience (at ~01:00 mark): memory only drops slightly (to ~2.66GB). When loading the second apartment, memory continues to increase until ~5GB, at which point the app crashes due to memory pressure. The issue is consistently visible under VM: IOSurface in Instruments. No leaks are detected. So it looks like RealityKit (or lower-level frameworks) keeps caching meshes and textures, and does not free them when RealityView is ended. But for my use case, these resources should be fully released once the ImmersiveSpace is dismissed, since new apartments will load entirely different models and textures. Cleanup Code Example Here’s a simplified version of the cleanup I’m doing: func clearAllRoomEntities() { for (entityName, entity) in entityFromMarker { entity.removeFromParent() if let modelEntity = entity as? ModelEntity { modelEntity.components.removeAll() modelEntity.children.forEach { $0.removeFromParent() } modelEntity.clearTexturesAndMaterials() } entityFromMarker[entityName] = nil removeSkyboxPortals(from: entityName) } entityFromMarker.removeAll() } extension ModelEntity { func clearTexturesAndMaterials() { guard var modelComponent = self.model else { return } for index in modelComponent.materials.indices { removeTextures(from: &modelComponent.materials[index]) } modelComponent.materials.removeAll() self.model = modelComponent self.model = nil } private func removeTextures(from material: inout any Material) { if var pbr = material as? PhysicallyBasedMaterial { pbr.baseColor.texture = nil pbr.emissiveColor.texture = nil pbr.metallic.texture = nil pbr.roughness.texture = nil pbr.normal.texture = nil pbr.ambientOcclusion.texture = nil pbr.clearcoat.texture = nil material = pbr } else if var simple = material as? SimpleMaterial { simple.color.texture = nil material = simple } } } Questions Is this expected RealityKit behavior (textures/meshes cached internally)? Is there a way to force RealityKit to release GPU resources tied to USDZ models when they’re no longer used? Should dismissing the ImmersiveSpace automatically free those IOSurfaces, or do I need to handle this differently? Any guidance, best practices, or confirmation would be hugely appreciated. Thanks in advance!

Hi, I am trying to load files from the Apple Vision Pro's storage into a Unity App (using Apple visionOS XR Plugin and not PolySpatial package). So far, I've tried using UnitySimpleFileBrowser and UnityStandaloneFileBrowser (both aren't made for the Vision Pro and don't work there), and then implemented my own naive file browser that at least allows me to view directories (that I can see from the App Sandbox). This is of course very limited: Gray folders can't be accessed, the only 3 available ones don't contain anything where a user would put files through the "Files" app. I know that an app can request access to these "Files & Folders": So my question is: Is there a way to request this access for a Unity-built app at the moment? If yes, what do I need to do? I've looked into the generated Xcode project's "Capabilities", but did not find anything related to file access. Any help is appreciated!

Hey Everyone, Happy New Year! I wanted to see if you have seen this before. I have added an attachment to the RealityView as a child on an entity that has a Billboard component set on it. I wanted to create the effect that the attachment is offset by .5 meters from center and follows the device as you move around it. IT works great until you try click a button. The attachment moves with the billboard, but the collision box around the attachment is not following it. If I position myself perfectly it works. Video Example: https://youtu.be/4d9Vx7K8MmU // // ImmersiveView.swift // Billboard Attachment // // Created by Justin Leger on 1/3/25. // import SwiftUI import RealityKit import RealityKitContent struct ImmersiveView: View { var rootEntity = Entity() var body: some View { RealityView { content, attachments in // Add the initial RealityKit content let sphereEntity = ModelEntity(mesh: .generateSphere(radius: 0.1), materials: [SimpleMaterial(color: .red, roughness: 1, isMetallic: false)]) sphereEntity.position = [0.0, 1.0, -2.0] let controlsPivotEntity = Entity() controlsPivotEntity.components[BillboardComponent.self] = .init() // Extract the attachemnt entity and disable it before its used. if let controlsViewAttachmentEntity = attachments.entity(for: PlacedThingControls.attachmentId) { controlsViewAttachmentEntity.position.z = 0.5 controlsPivotEntity.addChild(controlsViewAttachmentEntity) sphereEntity.addChild(controlsPivotEntity) } content.add(sphereEntity) } attachments: { Attachment(id: PlacedThingControls.attachmentId) { PlacedThingControls() } } } } #Preview(immersionStyle: .mixed) { ImmersiveView() .environment(AppModel()) } struct PlacedThingControls: View { static let attachmentId = "placed-thing-3D-controls" var body: some View { VStack { HStack(spacing: 0) { Button { print("🗺️🗺️🗺️ Map selected pieces") } label: { Text("\(Image(systemName: "plus.square.dashed")) Manage Mesh Maps") .fontWeight(.semibold) .frame(maxWidth: .infinity) } .padding(.leading, 20) Spacer() Button(role: .destructive) { print("🗑️🗑️🗑️ Delete selected pieces") } label: { Label { Text("Delete") } icon: { Image(systemName: "trash") } .labelStyle(.iconOnly) } .padding(.trailing, 20) } .padding(.vertical) .frame(minWidth: 320, maxWidth: 480) } .glassBackgroundEffect() } }

I'm trying to develop an immersive visionOS app, which you can move an Entity having a PerspectiveCamera as its child in immersive space, and render the camera view on 2D window. According to this thread, this seems to can be achieved using RealityRenderer. But when I added the scene entity loaded from realityKitContentBundle to realityRenderer.entities, I needed to clone all entities of the scene, otherwise all entities in the immersive space will disappear. @Observable @MainActor final class OffscreenRenderModel { private let renderer: RealityRenderer private let colorTexture: MTLTexture init(scene: Entity) throws { renderer = try RealityRenderer() // If not clone entities in the scene, all entities in the immersive space will disappear renderer.entities.append(scene.clone(recursive: true)) let camera = PerspectiveCamera() renderer.activeCamera = camera renderer.entities.append(camera) ... } } Is this the expected behavior? Or is there any other way to do this (move camera in immersive space and render its output on 2D window)? Here is my sample code: https://github.com/TAATHub/RealityKitPerspectiveCamera

Is there a way to provide CLHeading in visionOS so I can more accurately understand the direction in which a user is facing?

Hello Community, I'm encountering an issue with the latest iOS 17 update, specifically related to RoomPlan version-2. In iOS 16, when using RoomPlan version-1, we were able to display stairs in our app. However, after upgrading to iOS 17 and implementing RoomPlan version-2, the stairs are no longer visible. Despite thorough investigation, I couldn't find any option within the code to show or hide stairs, or any other objects for that matter. It seems like a specific issue with the update rather than a coding error on our part. Has anyone else encountered a similar problem? If so, I would greatly appreciate any insights or solutions you might have. It's crucial for our app functionality to have stairs displayed accurately, and we're currently at a loss on how to address this issue. Thank you in advance for any assistance you can provide. Best regards

In courses like Compose interactive 3D content in Reality Composer Pro Realitykit Engineers recommended working with Reality Composer Pro to create RealityKit packages to embed in our Realitykit Xcode projects. And, comparing the workflow to Unity/Unreal, I can see the reasoning since it is nice to prepare scenes/materials/assets visually. Now when we also want to run a Xcode Cloud CI/CD pipeline this seems to come into conflict: When adding a basic *.usdz to the RealityKitContent.rkassets folder, every build we run on Xcode cloud fails with: Compile Reality Asset RealityKitContent.rkassets ❌realitytool requires Metal for this operation and it is not available in this build environment I have also found this related forum post here but it was specifically about compiling a *.skybox.

Hi, I'm developing a prototype VisionOS game. How to access the bones or joints information when exporting a USD file from Blender to RCP? The animation in RCP works fine and the joints' information is correctly embedded in the USDA file (with usdchecker). However, RCP does not read it in USDA, USDC or USDZ. It should be possible based on Apple WWDC24 (Compose Interactive 3D content in RCP). I want to attach and detach an entity to a particular bone in certain moments. So I need the bones' data. They are standard mixamo animations. My mesh is a single unified mesh. Using Blender 4.4

I’m working with RealityView in visionOS and noticed that the content closure seems to run twice, causing content.add to be called twice automatically. This results in duplicate entities being added to the scene unless I manually check for duplicates. How can I fix that? Thanks.

After updating to the latest visionOS beta, visionOS 26 Beta 4 (23M5300g) the ‘Presenting images in RealityKit’ sample from the following link no longer builds due to an error. https://developer.apple.com/documentation/RealityKit/presenting-images-in-realitykit Expected / Previous: Application builds and runs on device, working as described in the documentation. Reality: Application builds, but does not run on device due to an error (shown in screenshot) “Thread 1: EXC_BAD_ACCESS (code=1, address=0xb)”. The application still runs on the simulator, but not on device. When launching the app from Xcode, it builds and installs correctly but hangs due to the respective error. When loading the app from the Home Screen, the app does not load, and immediately returns to the Home Screen. This Xcode project previously ran with no changes to code - the only change was updating the visionOS system software to the latest version. visionOS 26 Beta 4 (23M5300g) Is anyone else experiencing this issue?

What will be the pros and cons if unreal engine is used in apple vision pro? What are the issues that unreal will face in future on apple vision pro? What will be the specifications required for unreal to work on vision pro?

Hey there, I am working on an app that displays environmental data using PNG color channels to represent data ranges, which gets overlayed on a map. The sampled values aren't what I'm expecting though... for example an RGB value of 0x7f0000 (R = 0.5, G = 0, B = 0) would be seen as 0.21, 0, 0 in the shader. This basically makes it unusable if I'm trying to show scientific data... I'm half wondering if I am completely misunderstanding how sampling works in RealityKit / RealityComposerPro. Anybody have any idea why it works like this? Actual result (chart labels added in photoshop): Expected: Red > 0.1 Shader Graph

Hi, I have used the template code for Plane Detection and placing models on them from here https://developer.apple.com/documentation/visionos/placing-content-on-detected-planes This source code did not copy the animations in the preview model to the PlacedModel and hence I modified it to do a manual copy of animations and textures. There is a function called materialize() that does this and I was able to modify it to get it working where the placed models are now animating. The issue is when I apply gestures on them like drag or rotate. For those models that go through this logic I'm unable to add gestures even though I'm making sure that Collision and Input Target is set on the Placed Models. Has anyone been able to get this working or is it even a possibility? My materialize function func materialize() -> PlacedObject { let shapes = previewEntity.components[CollisionComponent.self]!.shapes // Clone render content first as we need its materials let clonedRenderContent = renderContent.clone(recursive: true) print("To be finding main model: \(descriptor.displayName)") // Find the main model in preview hierarchy func findMainModel(_ entity: Entity) -> Entity? { if entity.name == descriptor.displayName.replacingOccurrences(of: " ", with: "_") { print("Found main model: \(entity.name)") return entity } for child in entity.children { if child.name == descriptor.displayName.replacingOccurrences(of: " ", with: "_") { print("Found main model in children: \(child.name)") return child } } return nil } // Clone hierarchy preserving structure, names, and materials func cloneHierarchy(_ entity: Entity) -> Entity { print("Cloning: \(entity.name)") let cloned: Entity if let model = entity as? ModelEntity { // Clone with recursive false to handle children manually cloned = model.clone(recursive: false) if let clonedModel = cloned as? ModelEntity, let originalMaterials = model.model?.materials { // Preserve the original model's materials clonedModel.model?.materials = originalMaterials } } else { cloned = Entity() } // Preserve name and transform cloned.name = entity.name cloned.transform = entity.transform // Clone children for child in entity.children { let clonedChild = cloneHierarchy(child) cloned.addChild(clonedChild) } return cloned } print("=== Cloning Preview Structure ===") // Clone the preview hierarchy with proper structure let clonedStructure = cloneHierarchy(previewEntity) // Find and use the main model if let mainModel = findMainModel(clonedStructure) { print("Using main model for PlacedObject") let modelEntity: ModelEntity if let asModel = mainModel as? ModelEntity { print("Using asModel ") modelEntity = asModel } else { modelEntity = ModelEntity() modelEntity.name = mainModel.name // Copy children and transforms for child in mainModel.children { modelEntity.addChild(child) } modelEntity.transform = mainModel.transform } // Add collision component here let collisionComponent = CollisionComponent(shapes: shapes, isStatic: false, filter: CollisionFilter(group: PlacedObject.collisionGroup, mask: .all)) modelEntity.components.set(collisionComponent) // Create the placed object let placedObject = PlacedObject(descriptor: descriptor, renderContentToClone: modelEntity, shapes: shapes) // Set input target on the placed object itself placedObject.components.set(InputTargetComponent(allowedInputTypes: [.direct, .indirect])) return placedObject } else { print("Fallback to original render content") let placedObject = PlacedObject(descriptor: descriptor, renderContentToClone: clonedRenderContent, shapes: shapes) placedObject.components.set(InputTargetComponent(allowedInputTypes: [.direct, .indirect])) return placedObject } } My PlacedObject class where the init has the recursive cloning removed because it is handled in materialize class PlacedObject: Entity { let fileName: String // The 3D model displayed for this object. private let renderContent: ModelEntity static let collisionGroup = CollisionGroup(rawValue: 1 << 29) // The origin of the UI attached to this object. // The UI is gravity aligned and oriented towards the user. let uiOrigin = Entity() var affectedByPhysics = false { didSet { guard affectedByPhysics != oldValue else { return } if affectedByPhysics { components[PhysicsBodyComponent.self]!.mode = .static } else { components[PhysicsBodyComponent.self]!.mode = .static } } } var isBeingDragged = false { didSet { affectedByPhysics = !isBeingDragged } } var positionAtLastReanchoringCheck: SIMD3<Float>? var atRest = false init(descriptor: ModelDescriptor, renderContentToClone: ModelEntity, shapes: [ShapeResource]) { fileName = descriptor.fileName // renderContent = renderContentToClone.clone(recursive: true) renderContent = renderContentToClone super.init() name = renderContent.name // Apply the rendered content’s scale to this parent entity to ensure // that the scale of the collision shape and physics body are correct. scale = renderContent.scale renderContent.scale = .one // Make the object respond to gravity. let physicsMaterial = PhysicsMaterialResource.generate(restitution: 0.0) let physicsBodyComponent = PhysicsBodyComponent(shapes: shapes, mass: 1.0, material: physicsMaterial, mode: .static) components.set(physicsBodyComponent) components.set(CollisionComponent(shapes: shapes, isStatic: false, filter: CollisionFilter(group: PlacedObject.collisionGroup, mask: .all))) addChild(renderContent) addChild(uiOrigin) uiOrigin.position.y = extents.y / 2 // Position the UI origin in the object’s center. // Allow direct and indirect manipulation of placed objects. components.set(InputTargetComponent(allowedInputTypes: [.direct, .indirect])) // Add a grounding shadow to placed objects. renderContent.components.set(GroundingShadowComponent(castsShadow: true)) } required init() { fatalError("`init` is unimplemented.") } } Thanks

Hello, After watching the Work with Reality Composer Pro content in Xcode, I had created the following custom component. public struct TestComponent : Component, Codable{ public var text : String = "helloWorld" public init() {} } I had registered the custom component as suggested in App.init function init() { RealityKitContent.TestComponent.registerComponent() } The custom component is decoded and realityView shows the sphere, when I load the "Scene" from realityKitContent bundle. But if I export the scene to a separate file named "test_scene.usdz" on disk and shared to the simulator and then trying to load it load in reality view causes EXC_BAD_ACCESS #0 0x0000000194c8d508 in Swift._StringObject.getSharedUTF8Start() -> Swift.UnsafePointer<Swift.UInt8> () Printing the loaded entity, shows the customComponent but when trying to load in show realityview , crashes the app immediately. Is there a way to fix it?

Hey there, since SceneView has been marked as „deprecated“ for SwiftUI, I‘m wondering which alternatives should be considered for the following situation: I have a SwiftUI app (for iOS and iPadOS) where users can view (with rotate, scale, move gestures) 3D models (USDZ) in a scene. The models will be downloaded from web backend and called via local URL paths. What I tested: I‘ve tried ARView in .nonAR mode, RealityView, however I didn‘t get the expected response -> User can rotate, scale the 3D models in a virtual space. ARView in nonAR mode still shows the object like in normal AR mode without camera stream. I tried to add Gestures to the RealityView on iOS - loading USDZ 3D models worked but the gestures didn’t). Model3D is only available for visionOS (that would be amazing to have it for iOS) I also checked QuickLook Preview however it works pretty strange via Filepicker etc, which is not the way how the user should load the 3D models in my app. Maybe I missed something, I couldn’t find anything which can help me. I‘m pretty much stucked adopting the latest and greatest frameworks/APIs in my App and taking the next steps porting my app to visionOS. Long story short 😃: Does someone have an idea what is the alternative to SceneView for USDZ 3D models? I appreciate your support!! Thanks in advance!

I'm a novice in RealityKit and ARKit. I'm using ARKit in SwiftUI to show a cube with a number as shown below. import SwiftUI import RealityKit import ARKit struct ContentView : View { var body: some View { return ARViewContainer() } } #Preview { ContentView() } struct ARViewContainer: UIViewRepresentable { typealias UIViewType = ARView func makeUIView(context: UIViewRepresentableContext<ARViewContainer>) -> ARView { let arView = ARView(frame: .zero, cameraMode: .ar, automaticallyConfigureSession: true) arView.enableTapGesture() return arView } func updateUIView(_ uiView: ARView, context: UIViewRepresentableContext<ARViewContainer>) { } } extension ARView { func enableTapGesture() { let tapGestureRecognizer = UITapGestureRecognizer(target: self, action: #selector(handleTap(recognizer:))) self.addGestureRecognizer(tapGestureRecognizer) } @objc func handleTap(recognizer: UITapGestureRecognizer) { let tapLocation = recognizer.location(in: self) // print("Tap location: \(tapLocation)") guard let rayResult = self.ray(through: tapLocation) else { return } let results = self.raycast(from: tapLocation, allowing: .estimatedPlane, alignment: .any) if let firstResult = results.first { let position = simd_make_float3(firstResult.worldTransform.columns.3) placeObject(at: position) } } func placeObject(at position: SIMD3<Float>) { let mesh = MeshResource.generateBox(size: 0.3) let material = SimpleMaterial(color: UIColor.systemRed, roughness: 0.3, isMetallic: true) let modelEntity = ModelEntity(mesh: mesh, materials: [material]) var unlitMaterial = UnlitMaterial() if let textureResource = generateTextResource(text: "1", textColor: UIColor.white) { unlitMaterial.color = .init(tint: .white, texture: .init(textureResource)) modelEntity.model?.materials = [unlitMaterial] let id = UUID().uuidString modelEntity.name = id modelEntity.transform.scale = [0.3, 0.1, 0.3] modelEntity.generateCollisionShapes(recursive: true) let anchorEntity = AnchorEntity(world: position) anchorEntity.addChild(modelEntity) self.scene.addAnchor(anchorEntity) } } func generateTextResource(text: String, textColor: UIColor) -> TextureResource? { if let image = text.image(withAttributes: [NSAttributedString.Key.foregroundColor: textColor], size: CGSize(width: 18, height: 18)), let cgImage = image.cgImage { let textureResource = try? TextureResource(image: cgImage, options: TextureResource.CreateOptions.init(semantic: nil)) return textureResource } return nil } } I tap the floor and get a cube with '1' as shown below. The background color of the cube is black, I guess. Where does this color come from and how can I change it into, say, red? Thanks.

Reality Composer is no longer available in XCode 15 Release. Is this intended or will be available in later releases? Do I need to revert to Xcode15 Beta8 to get Reality Composer?

At WWDC25 we launched a new type of Lab event for the developer community - Group Labs. A Group Lab is a panel Q&A designed for a large audience of developers. Group Labs are a unique opportunity for the community to submit questions directly to a panel of Apple engineers and designers. Here are the highlights from the WWDC25 Group Lab for visionOS. I saw that there is a new way to add SwiftUI View attachments in my RealityView, what advantages does this have over the old way? Attachments can now be added directly to your entities with ViewAttachmentComponent. The removes the need to declare your attachments upfront in your RealityView initializer and then add those attachments as child entities. The new approach provides greater flexibility. Canyon Crosser and Petite Asteroids both utilize the new approach. ManipulationComponent looks really cool! Right now my app has a series of complicated custom gestures. What gestures does it handle for me exactly, and are there any situations where I should prefer my own custom gestures? ManipulationComponent provides natural interaction with virtual objects. It seamlessly handles translation and rotation. You can easily add manipulation to a SwiftUI view like Model3D with the manipulable view modifier. The new Object Manipulation API is great for most apps, and is a breeze to implement, but sometimes you might want a more custom feel, and that’s ok! Custom gestures are still fully supported for that scenario. I saw that there is a new API to also access the right main camera. What can I do with this? Correct, in visionOS 26, you can access the left and right main cameras. You can even access them simultaneously as a stereo pair. Camera access still requires a managed entitlement and an enterprise license, see Accessing the main camera for more details about those requirements. More computer vision and machine learning use-cases are unlocked with access to both cameras, we are excited to see what you will do! What do I need to do to add spatial accessory input for my app? First, use the GameController framework to establish a connection with the spatial accessory, and then listen for events from the controller. Then, you can use either RealityKit, ARKit, or a combination of both to track the accessory, anchor virtual content to it, and fine tune the accessory interaction with the content in your app. For more details, check out Discovering and tracking spatial game controllers and styli. By far, the most difficulty with implementing visionOS apps is SwiftUI window management…placing, opening, closing, etc. Are there any improvements to window management in visionOS 26? Yes! We recommend watching Set the scene with SwiftUI in visionOS. You can use the defaultLaunchBehavior to choose whether a particular window is presented (or suppressed) at launch. You can also prevent a window like a secondary toolbar from launching as the initial window using .restorationBehavior(.disabled). Adopting best practices for persistent UI provides a great overview of SwiftUI window management on visionOS. As for placing windows, there is still no API for an app to specify the placement of its windows other than relative placement. If that is a feature you are interested in, please file an enhancement request for it using Feedback Assistant! How to get access to the Enterprise API? First, request the entitlement and license through your Apple Developer or enterprise account. Once these have been granted, include the license and entitlement in your project. Then you can build, test, and distribute as an in-house app.

We're developing a VisionOS application, where we would like to do product recognition (like food items). We have enterprise entitlements and therefore also main camera access for VisionOS. We send this live camera frames to a trained CoreML model where we will receive 2D coordinates from the model detection prediction. Now, we would like to create a 3D anchor on the detected items so it can be visible for user. The 3D anchor is going to be the class name of the detected item. How do we transform this 2D coordinate from the model prediction to a 3D anchor?

Still don't understand why no one is clarifying about this Apple Video https://developer.apple.com/videos/play/wwdc2023/10111 At the end of this video, there's an incomplete tutorial about connecting a USDZ with mesh and Skeleton structure to the hand tracking system. No example project is linked, and no one is giving the community any clarification. Please can you help us to understand how to proceed?