API for turning regular photos into spatial photos?

With quite some excitement I read about visionOS 2's new feature to automatically turn regular 2D photos into spatial photos, using machine learning. It's briefly mentioned in this WWDC video:

https://developer.apple.com/wwdc24/10166

My question is: Can developers use this feature via an API, so we can turn any image into a spatial image, even if it is not in the device photo library?

We would like to download an image from our server, convert it on the visionPro on-the-fly and display it as a spatial photo.

Hello! We do not currently have a developer API for converting 2D photos into spatial photos using machine learning. Please do file a Feedback Assistant feature request describing your use-case.

Are 2D converted spatial photos permanent or temporary in the Photos app? That is, the converted spatial photos exported to other Apple devices also have depth information?

The conversion from 2D to spatial is saved as an edit to the original photo. This edit is synchronized to your iCloud Photo Library, including to other Apple devices. The converted photo contains both the original image, and the left- and right-eye images created during the conversion. This edit can always be reverted, to return to just original 2D image.

Note that the conversion process does not add a depth map to the image. Rather, the conversion generates and adds a pair of left- and right-eye images, derived from the original image.

API for turning regular photos into spatial photos?
 
 
Q