Apple Vision Pro Development Guide for Beginners

Apple Vision Pro runs visionOS — Apple's spatial operating system built from scratch for a new paradigm of computing. Unlike adapting an existing OS, visionOS was designed around the premise that people will interact with digital content floating in their physical space, using their eyes, hands, and voice as the only inputs.

Developing for Vision Pro means learning a new set of frameworks: RealityKit for 3D content, SwiftUI for UI, ARKit for world sensing, and Reality Composer Pro for spatial content creation. This guide covers the key concepts and practical steps to start building visionOS applications.


Development Environment Setup

- macOS Sequoia (14+) on Apple Silicon Mac (M1 or newer) recommended - Xcode 15+ with visionOS SDK: download from the Mac App Store - Apple Developer Program: $99/year — required to run on physical hardware and publish - Reality Composer Pro: included with Xcode, used for spatial content authoring - Swift language basics: SwiftUI tutorials at developer.apple.com are the best starting point ## visionOS App Types: Three Scene Experiences

Window (Shared Space)

The most familiar starting point. A Window app runs as a floating 2D panel in the user's space — like a Mac or iPad app that the user can position anywhere in their environment. SwiftUI Window Group powers this app type. All your existing SwiftUI knowledge applies directly.

Volume (Shared Space)

A Volume is a bounded 3D space rendered in the Shared Space alongside other apps. The user can place a Volume in their environment and see 3D content — a rotating Earth, a product visualization, a game board. RealityKit renders 3D content inside the Volume. This is the first truly spatial app type.

Full Space (Immersive Space)

An Immersive Space takes over the user's entire field of view — the equivalent of a VR experience. The user's real environment is hidden (or shown through passthrough). This app type enables fully immersive experiences, games, and spatial entertainment. Only one Immersive Space can be open at a time.

Core visionOS Frameworks

RealityKit

RealityKit is the 3D rendering and simulation framework for visionOS. It handles entity hierarchies (similar to Unity's GameObject/Component model), material rendering (physically based), spatial audio positioning, animation, and physics simulation. ModelEntity, AnchorEntity, and PointLightComponent are core building blocks.

ARKit on visionOS

ARKit on visionOS provides scene understanding: plane detection (floors, walls, ceilings, tables), mesh anchors (full room geometry), hand anchors (complete hand joint tracking), and image tracking. Use ARKit to anchor content to the real world, detect surfaces for physics, and respond to the user's spatial environment.

SwiftUI Integration

visionOS apps use SwiftUI for all 2D UI. The RealityView SwiftUI view bridges SwiftUI and RealityKit, allowing 3D content to be embedded within SwiftUI layouts. SwiftUI gestures are extended for spatial input — SpatialTapGesture, DragGesture with 3D coordinates, and RotateGesture3D.

Building Your First visionOS App

Open Xcode → New Project → visionOS → App. The template creates a Window-based app. Add a RealityView to display 3D content. Use Reality Composer Pro to create a .usda or .reality file with a 3D model. Load it in RealityKit via Entity.load(). Add a gesture to rotate the model on tap. Build and run in the Simulator.

The visionOS Simulator includes a virtual environment, simulated hands for testing gesture input, and the ability to test window positioning. It is excellent for development but cannot simulate eye tracking — test eye-based interactions on physical hardware.

Interaction Model: Eyes, Hands, Voice

Eyes select. Hands confirm. This is the core visionOS interaction model. A button is hovered when the user looks at it (system highlights it automatically). A pinch gesture with thumb and index finger confirms the selection. This model requires no deliberate pointing — the system tracks eye gaze automatically.

Designing for eye tracking means thinking about gaze ergonomics: interactive elements should be large enough to comfortably target, spaced to avoid accidental hovers, and positioned in the user's comfortable gaze arc (front 60° of their field of view). Text legibility in spatial environments requires minimum 16pt at comfortable reading distances.

Publishing to the App Store

visionOS apps submit to the App Store through the same process as iOS apps. In Xcode, archive your build, upload via Organizer, and submit through App Store Connect. visionOS apps are listed separately from iOS apps. The review process follows standard App Store guidelines.

The visionOS App Store is still early — competition is lower than iOS, and visibility for quality spatial apps is higher than it will be as the ecosystem matures. 2026 is an excellent time to ship a visionOS app and establish an early presence on the platform.