visionOS 27 + iOS 27 Object Tracking Demo
github.com/robomex/visionos-27-ios-27-object-tracking-demo- Category
- Developer Tools
- Rank
- No. 2033Tools index
Previous survey · No. 1892 ·
- Pricing
- Open Source
- Type
- TOOL
- Use case
- Research & Education
- GitHub
- 4 stars
- Date
About
This open-source demo pairs an iPhone running iOS 27 ARKit object tracking with a Vision Pro on visionOS 27, letting the iPhone place virtual objects on a table and the Vision Pro guide a user's hands to match real objects to those placements with their eyes closed. It showcases this year's high frame-rate tracking and more accurate, extended-mode-trained reference objects, fixing the lag that made visionOS 2's original object tracking unsuitable for real-time placement.
At a glance
This demo uses object tracking to align an iPhone and Apple Vision Pro around a shared physical object. It is worth studying if you want to see how two devices can agree on where virtual content belongs.
Good fit
Apple platform developers prototyping object anchors and shared spatial interactions.
Consider something else when
Skip it if you need a controlled comparison that isolates the effect of tracking frame rate.
Start here
Read the demo’s calibration steps and compare how each device uses the reference object. In a later trial, move it out of view and check what happens to the corresponding anchor.
Cost and requirements
MIT-licensed example. The repository lists Xcode 27, an iPhone on iOS 27 and an Apple Vision Pro on visionOS 27.
Limits to know
The author changed the training setup and frame rate between the compared examples. Treat device observations as leads to verify on your own hardware and target versions.
Sources checked 2026-11-05. Source-based assessment.
Why it made the leaderboard
Gives AR developers a working reference for the new visionOS 27/iOS 27 object-tracking APIs (high frame rate, extended-mode training, cross-device pose sharing) ahead of broader documentation catching up.
Tags
Comments (0)
No comments yet
Editorially curated, with community endorsements as a secondary signal. Corrections welcome.