iOS ScreenCaptureKit: background frame processing and capture scope

I’m evaluating ScreenCaptureKit for an explicitly user-initiated, on-device text-recognition workflow on iPhone.

Apple’s “Capturing screen content on iOS” sample currently describes iOS 27 full-display capture and capture continuing while the app is backgrounded. Before building a minimal prototype, I’d like to clarify:

  1. Can screen frames continue to be delivered and processed locally using Vision for text recognition while another app is foregrounded? What background configuration and processing limitations apply?
  2. Can capture on iPhone be restricted to a specific third-party app, or are the available scopes full-display capture and the capturing app’s own content?
  3. Is there a supported way to detect that the user has left the intended capture source and stop capture?
  4. Are there documented restrictions on capturing system apps, particularly Messages?

The intended configuration has explicit system authorization, a visible capture indication, stop/cancel controls, no audio or camera capture, no network transmission, and no retained screen recording. Temporary recognition results would expire and be discarded; only explicitly user-selected output would be retained.

This is a capability investigation, not a report of an observed failure. I haven’t run a device prototype yet.