How to blur faces in screenshots on a Mac
Short answer: cover each face so the person cannot be recognized, then save. A light blur looks tidy but leaves a face that people and software can still place, so the reliable move is to pixelate each face coarsely straight into the image. On a Mac there is no built-in blur, so you either improvise with a box or use a tool that scrubs the face into the picture. This guide covers when a face needs covering, why faces are a different problem from text, and exactly how to do it.
When a face actually needs covering
Most screenshots with a face in them are accidental. You grab a video-call window to show a layout problem and the shot is a grid of colleagues. You screenshot an admin panel for a bug report and a customer's profile photo is sitting in the corner. You capture something from a photo library and family or kids are in the frame. None of those people agreed to appear in the image you are about to send.
The same goes for anything social: a feed full of profile photos, a comment thread, a contacts list. The faces are not the point of the screenshot, they are just along for the ride. The considerate default is to cover any face that is not the reason you took the shot, unless you have a clear reason to show it and the okay to do so. It costs a few seconds and saves you from putting a stranger, a customer or a child online by accident.
Why faces are different from text
When you hide a code or an account number, the risk is that someone reads the exact characters back. With a face, the risk is recognition: a person does not need to reconstruct your screenshot pixel by pixel, they just need enough left to think "that's so-and-so". That is a much lower bar to clear, and it is why a gentle blur is not enough for a face.
A light blur softens the detail but keeps the shape, the hair, the skin tone and the pose, which is plenty for someone who already knows the person, and face-recognition systems handle soft, low-detail images well. So a blurred face can still be identified even when it looks anonymized at a glance. To actually remove recognition you have to remove the geometry of the face, not just take the edge off it. I have written up the underlying mechanics in my guide to how blurring and pixelating differ and what each leaves behind.
How each way of covering a face compares
Here is how the common ways of hiding a face stack up once you care about whether the person can still be identified, and whether the cover can be undone.
| Method | Can the person be identified? | Can it be undone? |
|---|---|---|
| Light blur | Often yes; the shape and features survive, and recognition software copes with soft images | A gentle blur can sometimes be partially reversed, the detail is still present |
| Emoji or sticker on top | No, while the sticker is in place | It is an overlay, so it can sometimes be moved or removed unless the image is flattened |
| Pixelate into the image | No; coarse blocks remove the geometry of the face entirely | No, the pixels are rewritten in place and there is nothing to recover |
The honest answer: scrub each face yourself
You might expect a face tool to detect faces for you. ScrubShot does not, on purpose. There is no face detection, no AI and no auto-detection of any kind. You scrub each face by hand with the Scrub tool, which pixelates it coarsely straight into the image.
That sounds like more work than letting a model find the faces, but auto-detection is exactly where faces get missed: someone half-turned away, a face in a reflection, a small one at the edge of a meeting grid, a child looking down. A missed face is the whole problem, because it is the one that ends up published. Doing it manually means a packed call grid with three or four faces costs you seconds, and nothing is left to a model that might skip one. Because the Scrub tool uses coarse pixelation rather than a soft blur, the geometry of each face is gone, not just softened.
The blocks it lays down are sampled at random from the region rather than averaged from the pixels they cover, so there is no overlay to lift off, no original face left in the file, and nothing for recovery software to work back from. None of this touches the network either, so a screenshot full of faces stays on your Mac the whole time, which matters most when those faces are not yours to share. The broader version of this thinking, applied to client work, is in my guide to anonymizing client screenshots without losing the credibility of a real shot.
How to blur faces on a Mac, step by step
This is the loop I use whenever a screenshot has people in it who do not need to be there.
- Press the keyboard shortcut. ScrubShot captures the screen under the cursor and opens the editor.
- Pick the Scrub tool and drag over each face in turn. Each one is pixelated straight into the image as you go, not added as a layer you have to flatten later.
- Work across the whole frame, not just the obvious faces: check reflections, the edges of a meeting grid, and any small profile photos. Use Undo if you overshoot.
- Copy the cleaned shot to the clipboard, or let it save to the ScrubShot folder in your Pictures, then paste it wherever it needs to go.
Because the pixels are rewritten in place, a scrubbed face cannot be un-scrubbed afterwards, and because the blocks are not derived from the face underneath, there is nothing to reconstruct. What an AI enhancer can invent from those blocks, and why the result is not the original person, is covered in can pixelation be reversed. If you want the wider picture of what each redaction method protects against, it is in my guide to redacting screenshots on a Mac.
FAQ
- Can a blurred face still be identified?
- A light blur often can be. The face is just softened, not removed, and the overall shape, hair, skin tone and pose still read clearly to a person who knows them, and face-recognition systems cope well with soft images. If the goal is that the person cannot be recognized, pixelate the face coarsely into the image so the geometry is gone, rather than gently blurring it.
- Is there a built-in way to blur a face on a Mac?
- Not as such. The Markup tools in Preview and Quick Look give you shapes, arrows and text but no blur or pixelate option, so people drop a solid box or an emoji over the face. A box works if you flatten the image afterwards, otherwise it stays an overlay that can be moved. For a face that genuinely cannot be recovered, pixelate it into the picture.
- Do I need permission to post a screenshot with someone’s face in it?
- This is judgement, not legal advice. People in a video call, a customer in an admin view or a child in a photo library never agreed to appear in your screenshot, so the considerate default is to cover faces unless you have a clear reason and the okay to show them. When in doubt, scrub the faces. It costs seconds and avoids putting someone online without their say.
- Does ScrubShot detect faces automatically?
- No, and that is deliberate. ScrubShot has no face detection, no AI and no auto-detection of any kind. You scrub each face yourself with the Scrub tool. Auto-detection sounds convenient but it misses faces at the edge of frame, in reflections or turned away, and a missed face is the whole problem. Doing it by hand means nothing is left to a model that might skip one.
Try it
ScrubShot is a Mac app that pixelates faces into the image so the people in your screenshots cannot be identified, all on-device. Press the shortcut, scrub out each face, then copy or save; the cleaned screenshot is the only version that ever leaves your Mac. There is a free 7-day trial with no card required. After that it is $30 once, with lifetime updates and no subscription.