Hey all, I'm working on a visionOS app that captures live frames from the left and right cameras of Apple Vision Pro using cameraFrame.sample(for: .left/.right). Apple provides documentation on encoding side-by-side frames into MV-HEVC spatial video using CMTaggedBuffer: Converting Side-by-Side 3D Video to MV-HEVC My question: Is there any way to render tagged frames (e.g. CMTaggedBuffer with .stereoView(.leftEye/.rightEye)) live, directly to a surface in RealityKit or Metal, without saving them to a file? I’d like to create a true stereoscopic (spatial) live video preview, not just render two images side-by-side. Any advice or insights would be greatly appreciated!

I've been struggling with this for far too long so I've decided to finally come here and see if anyone can point me to the documentation that I'm missing. I'm sure it's something so simple but I just can't figure it out. I can SharePlay our test app with my brother (device to device) but when I open a volumetric window, it says "not shared" under it. I assume this will likely fix the video sharing problem we have as well. Everything else works so smooth but SharePlay has just been such a struggle for me. It's the last piece to the puzzle before we can put it on the App Store.

In WWDC25 session What’s new for the spatial web, the presenter showed creating an immersive environment for a web page by adding to the page's HEAD section <link rel="spatial-backdrop" href="office.usdz" environmentmap="lighting.hdr"> My first attempt failed, and I am trying to track down why. Before I search all the potential failure paths, I wanted to ask the community, Is this feature available in the latest visionOS 26 beta? I haven't seen anyone talk about their use of the feature yet.

Hello, I was wondering how I can initialize an ImageAnchoringSource using https://developer.apple.com/documentation/realitykit/anchoringcomponent/imageanchoringsource/init(_:) When I construct one using a URL, it doesn't seem to be tracked and I see in the following when I debug print the component: ▿ 0 : AnchoringComponent ▿ target : Target ▿ referenceImage : 1 element ▿ from : ImageAnchoringSource ▿ url : Optional<URL> ▿ some : file:///var/mobile/Containers/Data/Application/D1126EA0-A1D7-468F-A40C-8578B7F5BDDF/Library/Caches/CodeCache/0E457AA7-2195-48B9-9DD4-58CEB9397F69.png - _url : file:///var/mobile/Containers/Data/Application/D1126EA0-A1D7-468F-A40C-8578B7F5BDDF/Library/Caches/CodeCache/0E457AA7-2195-48B9-9DD4-58CEB9397F69.png - _parseInfo : nil - _baseParseInfo : nil - name : nil - group : nil ▿ trackingMode : TrackingMode - trackingMode : 2 Is there a specific format for the parseInfo? When I use the same image to make an image anchoring source by group and name in AR Resources, it is tracked. Thank you!

When building a multiplayer Tabletop game, the documentation includes how to attach a custom TabletopNetworkSessionCoordinator, which could be used in addition to TabletopGame.MultiplayerDelegate. But so far, we have been unable to create these types of custom coordinators or have a delegate that works. Our current setup with our generic GroupActivity works by sending the session to TabletopGame's coordinateWithSession method (like in the current sample project), but we didn't find a way to access and control, for example, the arbiter, seats, player events, among other features mentioned on https://developer.apple.com/documentation/tabletopkit/tabletopnetworksession. Is correct to expect having access to the participants, messenger, or journal without having to maintain a parallel coordinator?   possibly we are missing something here; any suggestions?

With Xcode 26, loading ressources with RealityKit is extremely slow. Here my project takes almost 50 seconds to load. I also get multiple Hang detected messages in the console: When I uncheck "Debug executable" in the schema, the same project loads in 2 seconds. I'm using RealityKit asynchronous loading: private static func loadFromRealityComposerPro( named entityName: String, fromSceneNamed sceneName: String ) async -> Entity? { var entity: Entity? do { let scene = try await Entity( named: sceneName, in: visionPetsContentBundle ) entity = scene.findEntity(named: entityName) } catch { print( "Error loading \(entityName) from scene \(sceneName): \(error.localizedDescription)" ) } return entity } Anyone having the same problem?

Hi everyone, I’m working with RealityKit on visionOS and I’m seeing unexpected behavior when the user long-presses the Digital Crown, which recenters the world. Observed behavior: When the world is recentered via long-pressing the Crown, the models remain visually in the correct place (as expected). However, if I query the model’s position or transform immediately after recentering (e.g. entity.position or similar), I still get the old values from before recenter. As soon as I interact with the model using a gesture (drag/rotate/scale), the position updates and then querying it returns the correct, updated values. So effectively: Recenter happens Visual position is correct Programmatic position remains stale First gesture causes the position to “snap” to the correct updated value Questions: Is there any event, notification, or callback that fires when the world is recentered due to a long press of the Crown button? Is there a recommended way to get the updated world-space transform immediately after recenter, without waiting for a gesture? Is this expected behavior due to deferred/lazy transform updates in RealityKit? Right now it feels like recentering updates the coordinate system but doesn’t immediately commit new transform values to entities until some interaction occurs. Any guidance or best-practice patterns for handling this would be appreciated. Thanks!

Has Roomplan been abandoned? Two years have gone by without comments from Apple on improvements. Are the improvements behind the scenes? Is there going to be any major updates?

Hi, I'm trying to correct the lens distortion in frames provided by Enterprise API camera frame provider. The frames provided seem to have only in/extrinsics info, but not the distortion lookup table. Is there some magic setting, or function to do that (I can't seem to find anything like this)? Or is there a way to use AVCameraCalibrationData together with provider?

We can add ornaments to popovers shown by PresentationComponent, but I’m not sure if we should. While working on the editor for entities in a Volume-based app, I had the idea to add ornaments to the presented views. The entire app exists inside a volume. A user can tap a item to present a popoverUI to edit it. This is displayed using the new PresentationComponent in visionOS 26. Ornaments have a new attachment anchor option this year: .parent(). .ornament(attachmentAnchor: .parent(.top), ornament: {...}) This works well in the Simulator. We can add ornaments around this popover view just like we would with a window. Unfortunately, when I run this on device I get a different experience. Any part of the ornament that overlaps with the popover content isn’t rendered correctly. Sometimes it entirely disappears, other times it becomes partially transparent. We could use content alignment to try to make sure the ornament doesn’t overlap the popover content. .ornament(attachmentAnchor: .parent(.top), contentAlignment: .bottom, ornament: {...}) This works sometimes–but not all the time. It’s not clear if this is a bug or not, because I’m not sure if we are even supposed to be able to use ornaments in this way. Here is my hierarchy: An app opens as a Volume Volume presenting a RealityView, with its own ornament using .scene() anchor Multiple Entities with Presentation Component show an edit view The view uses .parent() anchor to add ornaments. What makes me unsure is that other methods for drawing UI in RealityView don’t seem to work with ornaments. For example, if I add an attachment to show a view with the ornament–even when I use the .parent() anchor–the ornament is anchor to the volume, not the attachment view. So what do we think? Is this a rendering bug? Are ornaments intended to work with attachments and presentations?

.glassEffect(.regular, in: .rect(cornerRadius: 24)) error; 'glassEffect(_:in:isEnabled:)' is unavailable in visionOS This is not surprising since visionOS already has a native glass interface that formed a model for the other OS's, but this error will create additional overhead for developers creating multi-platform apps that include visionOS.

I want to open the control center in Vision Pro’s Xcode simulator. Can I open it? If I can, please tell me how to do it. Thank you.

If I long press on an element, the sidebar disappears and then a Done appears on the screen, but nothing else changes, so what are the Environments in Vision Pro's Simulator?

Using Xcode v26 Beta 6 on macOS v26 Beta 25a5349a When pressing on the home button of the visionOS simulator, I am not positioned in the middle of the room like would normally be. This occurred when moving a lot in the space to find an element added to an ImmersiveSpace. How to resolve: restart simulator device. See attached the pictures of the visionOSSimulatorCorrectHomePosition and the visionOSSimulatorMisallignedHomePosition.

Hi, I'm working with CameraFrameProvider from Enterprise API. Is it always capped at 30fps, or is there something I can switch to get more? I assume it is capped at 30, so let me cram in additional question here :). If I'd get a developer strap and attach an external camera capable of doing >30fps, will I get the full stream, or some other limitation will kick in?

At the moment the map kit APls only support non-volumetric maps (i.e. in a window or in a volume, but on a 2D surface). Is support for 3D volumetric maps in VisionOS in the works? And if so when can we expect it to be available?

The landing page for visionOS 26 mentions The Unified Coordinate Conversion API makes moving views and entities between scenes straightforward — even between views and ARKit accessory anchors. This WWDC session very briefly shows a single example of using this, but with no context. For example, they discuss a way to tell the distance between a Model3D and an entity in a RealityView. But they don't provide any details for how they are referencing the entity (bolts in the slide). The session used the BOT-anist example project that we saw in visionOS 2, but the version on in the Sample Code library has not been updated with these examples. I was able to put together a simple example where we can get the position of a window relative to the world origin. It even updates when the user recenters. struct Lab080: View { @State private var posX: Float = 0 @State private var posY: Float = 0 @State private var posZ: Float = 0 var body: some View { GeometryReader3D { geometry in VStack { Text("Unified Coordinate Conversion") .font(.largeTitle) .padding(24) VStack { Text("X: \(posX)") Text("Y: \(posY)") Text("Z: \(posZ)") } .font(.title) .padding(24) } .onGeometryChange3D(for: Point3D.self) { proxy in try! proxy .coordinateSpace3D() .convert(value: Point3D.zero, to: .worldReference) } action: { old, new in posX = Float(new.x) posY = Float(new.y) posZ = Float(new.z) } } } } This is all that I've been able to figure out so far. What other features are included in this new Unified Coordinate Conversion? Can we use this to get the position of one window relative to another? Can we use this to get the position of a view in a window relative to an entity in a RealityView, for example in a Volume or Immersive Space? What else can Unified Coordinate Conversion do? Are there documentation pages that I'm missing? I'm not sure what to search for. Are there any Sample projects that use these features? Any additional information would be very helpful.

Hi everyone, I’m building a visualization app for VisionPro that uses SharePlay and GroupActivities to explore datasets collaboratively. I’ve successfully implemented the new SharedWorldAnchor feature, and everything works well with nearby, local participants. However, I’m stuck on one point: How can I share a world anchor with remote participants who join via FaceTime as spatial personas? Apple’s demo app (where multiple users move a plane model around) seems to suggest that this is possible. For context, I’m building an immersive app with Metal rendering. Any guidance or examples would be greatly appreciated! Thanks, Jens

I have been concentrating on developing the visionOS application. While I am currently quite familiar with RealityKit, CompositorServices has also captured my attention. I have not yet acquired knowledge of CompositorServices. Could you please clarify whether it is essential for me to learn CompositorServices? Additionally, I would appreciate it if you could provide insights into the advantages of RealityKit and CompositorServices.

I'm trying to run a PhotogrammetrySession based on photos taken in an AVCaptureSession and stored as .heic files. When I load the files I'm always seeing the error "Sample 0 missing LiDAR point cloud!" showing up for each individual sample. Debugging shows that sample.depthDataMap is populated, also the .heic contains depth data which can be extracted using e.g. heif-convert on my Mac. Comparing the .heic I created to one of the ObjectCaptureSession which doesn't show the LiDAR warning, I noticed the only difference being the HEIC information here: So my questions are: Are these the missing information in my manual capture causing this warning? Can I somehow add these information in an AVCaptureSession? Do these information allow better photogrammetry results?