VisionMeshCapture is a spatial mesh capture and analysis tool for Apple Vision Pro. It allows users to define a 3D capture box in immersive space, preview reconstructed environment geometry, tune capture quality in real time, inspect mesh color information where available, align imported 3D models, and export structured mesh packages including PLY, USDZ, CSV, and JSON files.

The app is designed for professional, educational, research, visualization, spatial documentation, ergonomic study, design review, and technical prototyping workflows. It helps users capture and organize local 3D spatial information from the surrounding environment and imported assets.

Key Features

Immersive selection-box capture
Place a visible yellow 3D capture box in immersive space. The box includes a yellow perimeter, always-visible center sphere, and draggable transparent surface handles for resizing and refinement.

Real-time capture preview
Preview the reconstructed faces inside the selection box before committing the final capture. The HUD displays live diagnostics such as source faces, box hits, filtered faces, retained faces, rejected faces, triangle area range, and nearest reconstructed surface distance.

Automatic capture tuning
The app can automatically choose practical capture settings to reduce empty-box results and improve successful mesh extraction. Users can also manually adjust resolution, face budget, minimum triangle area, preview opacity, and color-blending settings.

Enterprise camera-aware color path
When approved enterprise main-camera access is available, the app can use camera-frame information to improve color sampling and vertex-color export. If enterprise access is unavailable, the app remains fully functional using the scene-reconstruction mesh and classification/fallback color path. Apple describes Vision Pro main-camera access as part of enterprise APIs, so this capability should be enabled only with the approved entitlement and license.  

Scene reconstruction and world sensing
The app uses visionOS world-sensing / scene-reconstruction capabilities to understand surrounding geometry inside the selected capture area. Apple requires a world-sensing privacy usage description for features such as scene reconstruction, plane detection, image tracking, and related ARKit data access.  

Multiple USDZ loading
Load multiple USDZ models into immersive space. Each loaded model keeps its own stored transform and can be placed at the center of the capture box.

Point-cloud alignment
Use captured points inside the selection box as an input point cloud and align the last loaded USDZ model to the captured mesh using point-cloud matching techniques.

Mesh and color analytics
For imported vertex-colored meshes, the app can read vertex positions and color values, allowing export of structured (x, y, z, r, g, b, a) information for analysis, clustering, thresholding, review, or post-processing.

Export package
Export structured capture packages containing:

* captured environment PLY
* captured environment USDZ
* imported USDZ models
* captured vertices CSV
* captured faces CSV
* color/mesh analysis CSV
* manifest JSON
* package metadata

 

Professional workflow
The app includes a scrollable control window, collapsible menus, immersive HUD, progress bars, export status, help menu, privacy text, terms of use, and a first-launch consent page.

Example Uses

VisionMeshCapture may be useful for:

* spatial documentation of local environments
* design and architecture review
* object-to-environment alignment testing
* mesh reconstruction experiments
* ergonomic and posture-related spatial research
* technical prototyping for spatial computing
* education and demonstration of 3D capture concepts
* comparison of imported USDZ models against captured geometry
* point-cloud registration experiments
* exporting local mesh data for external review

Important Limitations

VisionMeshCapture is not a medical device and does not provide diagnosis, treatment recommendations, surgical planning, or automated clinical decisions. Any measurements, reconstructions, color samples, alignments, or exported files are intended for visualization, research, education, technical review, and documentation support only.

The app does not replace professional judgment. Accuracy depends on device conditions, lighting, reconstruction quality, surface texture, camera availability, user-selected thresholds, and environmental factors.