We’ve all taken a photo that would’ve been perfect if only the camera had been positioned a little differently, right? Well, Apple now wants to fix that with Spatial Reframing, a new Apple Intelligence feature that can virtually reposition the camera after the shot is taken while using AI to naturally fill in the missing parts of the scene.
Spatial Reframing uses AI to virtually move the camera after the shot
According to Apple, Spatial Reframing combines on-device spatial models, built using technology developed for Apple Vision Pro, with its image generation models running on Private Cloud Compute. Users can simply drag around a photo during editing, and the perspective shifts as though the camera itself had been moved while capturing the original scene.
Apple says only the newly exposed portions of the image are generated by AI, while the original content remains untouched. The company demonstrated the feature by repositioning a family photo, moving the framing lower to create a more balanced composition while naturally filling in the missing background around the edges.
Alongside Spatial Reframing, Apple also announced an upgraded Extend tool that can expand images to create more breathing room or straighten horizons without aggressively cropping the original shot.
This might be the closest thing to a photographic time machine yet
Unlike traditional AI photo editing that simply adds or removes objects, Spatial Reframing tries to recreate a shot you never actually took. That’s a far more ambitious idea, because it asks AI to understand the three-dimensional layout of a scene before generating only the missing pieces needed to support a new perspective.
If it works as well in the real world as it did on stage, this could easily become one of the most useful Apple Intelligence features announced so far. After all, everyone has taken a badly framed photo at some point, and Apple is essentially promising a second chance without asking anyone to go back in time.
