Covers the three scene types available — WindowGroup (2D windows in space), Volume (3D bounded content), and ImmersiveSpace (full immersive experience) — with references on windows/volumes/spaces, RealityView/3D content, and ornament-based 2D UI attachments.

Best practices: start with familiar 2D windows before adding depth, introduce volumes gradually for 3D content, use ornaments to attach 2D UI to 3D scenes, respect the user's physical space rather than overwhelming it, design for natural hand tracking, and avoid rapid movements that cause eye discomfort.

visionOS Platform

visionOS-specific development for Apple Vision Pro spatial computing.

Agent Workflow (MANDATORY)

Before ANY implementation, spawn 3 agents in parallel, one Agent call each with a name:

1. fuse-ai-pilot:explore-codebase - Analyze existing visionOS patterns
2. fuse-ai-pilot:research-expert - Verify latest visionOS 27 docs via Context7/Exa
3. mcp__apple-docs__search_apple_docs - Check spatial computing patterns