What does a screen recording capture?

By Check Before Send · Updated

Whatever is on the screen while it runs, and sound too if you turn the microphone on. Apple describes an iPhone screen recording as capturing what is happening on your screen and saving it as a video, with optional audio. Anything that appears while it records can end up in the video.

This guide covers what Apple documents, what can show up in a recording, and how well our own tool found invented private details in a test recording.

What Apple says

Apple's guide to screen recordings says you can make a screen recording to capture what is happening on your iPhone screen and save it as a video, and that you can turn on the microphone to add audio. It says recordings are saved to your photo library, where they appear under a Screen Recordings media type.

Apple's page does not list which on-screen items appear in a recording, and it does not say how notifications behave. We did not test that on a phone.

What can show up

  • An email address, phone number or account number on any screen you open.
  • A web address or a code in a browser bar.
  • A notification banner that arrives from another app, if the recording captures it.
  • Your own voice, and other people's, if the microphone is on.
  • The details of the file, saved where the video is saved.

The UK data protection regulator's guidance on footage says that if individuals can be identified from a recording, it is personal data about them, and that redaction of visual and audio data may be needed before sharing.

What we tested

We made a four-second test screen recording from a made-up support page with invented data: an email address, a phone number, an IP address, a web address, and a session token that is on screen for only half a second.

How the text was readFound
One pass over the whole frame, normal size1 of 5
Whole frame enlarged twice3 of 5
Whole frame enlarged three times4 of 5
The tokenNot found by any whole-frame pass, because the highlighted chip was ignored; reading that area on its own located it but misread the characters

So we read text at three times the size, and then look at regions too. A short item matters: a half-second item appears in about 12 frames of a 25-frames-per-second video, so looking at every 10th frame catches it and every 25th does not.

Reading text this way is slow, and it can misread, so treat what it finds as a list to confirm rather than a promise that everything was read. For videos you export from our page, we do not run a second scan of the picture for faces or plates afterwards. Watching the result is your check.

What to do

  1. Before you record, turn off or hide anything you do not want captured, including notifications.
  2. After you record, watch the whole recording, especially moments when you switch apps.
  3. Cover what you find, and check the exported copy again, start to end.

Check a screen recording on this site

Our page runs in your browser and does not upload the video; the Trust Center explains how to confirm that. It looks for readable text, faces and plates, and exports a new file with the areas you choose covered. For the file's own details, see does a video file carry location and date, or browse the other guides.

Sources