# Build a Spatial Video Recording Camera UI in SwiftUI > By Lawrence Arya, Founder & CEO of VP0. Published 2026-06-08. 6 min read. > Source: https://vp0.com/blogs/spatial-video-recording-ui-clone-swiftui **TL;DR.** A spatial video recording UI is a camera screen for capturing stereoscopic 3D video, the kind that plays back with real depth on a Vision Pro. The build sits on an AVFoundation capture session configured for the spatial format, but the real work is detecting which devices support it and coaching the shot: landscape, steady, well-lit, at a comfortable distance, because stereo capture is unforgiving. Start from a camera template so the preview, the record control, the guidance, and the states are already shaped. ## What spatial video recording actually is Spatial video is stereoscopic video: two slightly offset views, one per eye, captured together so it plays back with real depth on a [Vision Pro](https://en.wikipedia.org/wiki/Apple_Vision_Pro). On iPhone it is recorded with two rear cameras working as a stereo pair, at roughly 1,920 by 1,080 per eye, and the result is a normal-looking clip on a flat screen that becomes three-dimensional in a headset. So a spatial video recording UI is a camera screen with a specific job: capture a clean stereo pair and guide the user into the shots that actually look good in 3D. The honest starting point is hardware. Spatial video records only on the devices that have the camera arrangement for it, currently recent iPhone Pro models and the Vision Pro itself, so the recording screen has to detect support and say so rather than offer a button that does nothing. ## The capture, and why the guidance matters Under the UI, capture runs through [AVFoundation](https://developer.apple.com/documentation/avfoundation), configured for the spatial format on a supported device. The recording itself is not the hard part; the guidance is. Stereo capture is unforgiving in ways flat video is not: it wants landscape orientation, steady hands, good light, and subjects at a comfortable distance, because depth falls apart with camera shake, harsh shadows, or anything too close. A spatial recording screen that ignores this produces clips that feel wrong in the headset, so the UI earns its keep by coaching the shot in real time. That coaching is mostly small, honest prompts: a nudge to rotate to landscape, a steadiness hint, a note when the light is poor. They are the difference between footage that lands in 3D and footage that gives people eye strain. ## The recording screen, piece by piece The screen itself is a familiar camera layout with spatial-specific touches. There is the live preview, a clear spatial badge so the user knows they are in stereo mode, a record button with a running timer, and a toggle between spatial and standard capture for when the moment does not need depth. Around those sit the guidance prompts and a quick path to review the last clip. The states matter: a not-supported state on older hardware, a recording state, and a saved state that hints the clip is best viewed on a Vision Pro. The broader headset surface is covered in [Apple Vision Pro React XR components](/blogs/apple-vision-pro-react-xr-components/), and the camera-with-overlay pattern parallels a [SwiftUI ARKit camera overlay](/blogs/swiftui-camera-arkit-overlay-code/). ## Building it from a template The preview, the spatial badge, the record control, the guidance prompts, and the supported-versus-not states are the same in every spatial recorder, so they are worth starting from. A free [VP0](https://vp0.com) design ships the camera screen, the record control, the guidance overlay, and the states as a SwiftUI file with a machine-readable source page, so pasting the link into Claude Code or Cursor gives the agent the recording UI to wire to the capture session. The spatial-placement thinking continues in an [AR object placement target](/blogs/ar-object-placement-target-ui-swiftui/), where depth and position are also the product. ## Common mistakes building a spatial recorder The recurring ones come from ignoring the format's constraints. Offering a spatial record button on unsupported hardware confuses users, so detect support first. Skipping the orientation and steadiness guidance produces clips that feel wrong in 3D. Treating spatial like flat video, no depth-friendly framing, wastes the whole point. Forgetting the not-supported and saved states leaves the screen feeling broken or unclear about where to watch the result. And burying the standard-capture toggle forces depth on moments that do not need it. ## Key takeaways: a spatial video recording UI - **Spatial video is a stereo pair.** Two offset views captured together, three-dimensional on a Vision Pro. - **Hardware gates it.** Detect support and show an honest not-supported state instead of a dead button. - **Guidance is the real feature.** Landscape, steadiness, light, and distance coaching make footage that works in 3D. - **The screen is a camera with spatial touches.** Preview, spatial badge, record timer, and a spatial-or-standard toggle. - **Start from a template.** A free VP0 SwiftUI design gives an agent the recording UI and states to wire to the capture session. ## Frequently asked questions **How do I build a spatial video recording UI in SwiftUI?** Build a camera screen on top of an AVFoundation capture session configured for the spatial format, but first detect whether the device supports spatial capture and show an honest not-supported state if it does not. The UI is a live preview with a spatial badge, a record button and timer, a spatial-or-standard toggle, and real-time guidance for landscape orientation, steadiness, light, and distance, because stereo capture is unforgiving. A free template gives you the screen, the record control, and the states to start from. **What is the safest way to build this with Claude Code or Cursor?** Give the agent the recording-screen template and let it wire the capture session, while you handle the device-support detection. A free VP0 SwiftUI design has a machine-readable source page with the preview, the spatial badge, the record control, the guidance overlay, and the supported-versus-not states, so Claude Code or Cursor builds against a real screen. That avoids the common result where an AI tool ships a spatial record button that does nothing on unsupported hardware and skips the guidance that makes 3D footage usable. **Can VP0 provide a free SwiftUI template for a camera or spatial recording screen?** Yes. VP0 has free camera and recording designs in SwiftUI with the preview, the record control, the guidance overlay, and the capture states already built, each exposing an AI-readable source page. Because the screen exists, your agent connects it to an AVFoundation capture session instead of reinventing the recording UI and the state handling that usually trip up hand-built camera screens. **What devices can record spatial video?** Spatial video records on devices with the camera arrangement for stereo capture, currently recent iPhone Pro models and the Apple Vision Pro, and it plays back in three dimensions on the Vision Pro. On unsupported hardware there is no spatial capture, which is why a good recording UI detects support and shows a clear not-supported state rather than offering a button that cannot work. The clip still looks normal on a flat screen and only reveals its depth in the headset. **What common errors happen when vibe coding a spatial recorder?** Offering a spatial record button on hardware that cannot do it, skipping the orientation and steadiness guidance, and framing shots like flat video are the frequent ones. Forgetting the not-supported and saved states leaves the screen unclear, and hiding the standard-capture toggle forces depth on moments that do not need it. Detect support first, coach the shot in real time, and make it obvious the result is best viewed on a Vision Pro. ## Frequently asked questions ### How do I build a spatial video recording UI in SwiftUI? Build a camera screen on top of an AVFoundation capture session configured for the spatial format, but first detect whether the device supports spatial capture and show an honest not-supported state if it does not. The UI is a live preview with a spatial badge, a record button and timer, a spatial-or-standard toggle, and real-time guidance for landscape orientation, steadiness, light, and distance, because stereo capture is unforgiving. A free template gives you the screen, the record control, and the states to start from. ### What is the safest way to build this with Claude Code or Cursor? Give the agent the recording-screen template and let it wire the capture session, while you handle the device-support detection. A free VP0 SwiftUI design has a machine-readable source page with the preview, the spatial badge, the record control, the guidance overlay, and the supported-versus-not states, so Claude Code or Cursor builds against a real screen. That avoids the common result where an AI tool ships a spatial record button that does nothing on unsupported hardware and skips the guidance that makes 3D footage usable. ### Can VP0 provide a free SwiftUI template for a camera or spatial recording screen? Yes. VP0 has free camera and recording designs in SwiftUI with the preview, the record control, the guidance overlay, and the capture states already built, each exposing an AI-readable source page. Because the screen exists, your agent connects it to an AVFoundation capture session instead of reinventing the recording UI and the state handling that usually trip up hand-built camera screens. ### What devices can record spatial video? Spatial video records on devices with the camera arrangement for stereo capture, currently recent iPhone Pro models and the Apple Vision Pro, and it plays back in three dimensions on the Vision Pro. On unsupported hardware there is no spatial capture, which is why a good recording UI detects support and shows a clear not-supported state rather than offering a button that cannot work. The clip still looks normal on a flat screen and only reveals its depth in the headset. ### What common errors happen when vibe coding a spatial recorder? Offering a spatial record button on hardware that cannot do it, skipping the orientation and steadiness guidance, and framing shots like flat video are the frequent ones. Forgetting the not-supported and saved states leaves the screen unclear, and hiding the standard-capture toggle forces depth on moments that do not need it. Detect support first, coach the shot in real time, and make it obvious the result is best viewed on a Vision Pro. --- *Published on the [VP0 Journal](https://vp0.com/blogs). Free to read, index and cite with attribution.*