How to Blur or Pixelate Part of a Screenshot Before Sharing It
Blur, pixelate or black out sensitive parts of a screenshot on Windows, Mac or in the browser — and why painting over text with a marker isn't redaction.

The screenshot was exactly right — bug framed, URL bar visible, error message in play. Except your customer's email address is sitting in the corner of it, and the Slack channel that needs this image needs it now.
Blurring part of a screenshot takes ten seconds once you know which tool to reach for. The trap is that the obvious fix — scribbling over the sensitive part with a marker or highlighter — is not a fix at all. That stroke is often semi-transparent, and the text under it can be brought back with nothing more than a brightness slider.
This guide covers what actually counts as redacted, the honest state of the built-in tools on Windows, macOS and mobile (spoiler: none of them blur), a browser-first way to pixelate web screenshots at capture time, and what to scan for before you hit paste.
Key takeaways#
- Paint-over is not redaction: the marker and highlighter in iOS Markup, Android Markup and Windows Snipping Tool draw semi-transparent strokes — anyone can recover the text underneath with brightness or levels adjustments.
- Pixelation can be reversible too: the Depix tool demonstrated recovering pixelated passwords from screenshots, so for secrets use coarse mosaic, a solid bar, or cropping — never a light blur.
- Destruction beats scrambling: an opaque black bar or a crop removes the pixels entirely; blur only rearranges them. The safest redaction is the one that destroys data.
- Windows and macOS have no built-in blur: Snipping Tool and Preview offer pens and shapes, not blur. An opaque filled rectangle is the honest built-in answer on both.
- Redact locally, not in the browser tab of a website: online blur tools are the easiest to find and the easiest to regret — most work by uploading the exact image you are trying to protect.
Table of contents#
- What actually counts as redacted
- Method 1: Blur or pixelate before the screenshot leaves the page
- Method 2: On Windows, cover it — the built-in tools don't blur
- Method 3: On macOS, use an opaque shape in Preview
- Method 4: On iPhone and Android, skip the marker
- Method 5: Online blur tools — know what you're uploading
- Which method fits which situation
- What to redact before you share
- FAQ
What actually counts as redacted#
Before the tool talk, the part most tutorials skip: redaction fails in specific, documented ways. Three of them cover almost every case.
1. The paint-over that isn't. In 2018, 9to5Mac published a PSA showing that text "blacked out" with the marker in iOS Markup can be revealed again by adjusting brightness in the editing tools — the marker stroke is semi-transparent by design, so it darkens the text without replacing the pixels. The same class of failure applies to the highlighter in Windows Snipping Tool and the sketch strokes in macOS Preview. If the redaction is a layer of ink sitting on top of the text, assume it can be peeled off.
2. Pixelation that un-pixelates. In 2020, a researcher released Depix, a tool that recovers text from screenshots pixelated with the common linear box filter — it brute-matches what each pixelated character block would look like and rebuilds the original string. Forbes covered it when it landed, and JumpSEC's later analysis explains the matching technique in detail. The demonstration targeted monospaced text — passwords in a terminal — but the lesson generalizes: coarse mosaic destroys information, fine pixelation merely reorganizes it.
3. Blur is cosmetic. Gaussian blur spreads each pixel's information into its neighbors. Cranking the radius makes recovery impractical for the human eye, but the information is still statistically present in the file — and deconvolution research keeps nibbling at the edges. Blur is a presentation choice, not a cryptographic one.

Only the last two destroy the pixels — everything above them can be undone with filters.
⚠️ The rule that survives all three failures: redact by destroying pixels — a solid opaque bar, a crop, or mosaic coarse enough that several characters collapse into each block — and flatten the result into a new image. Treat anything that stays a separate layer on top of the text as recoverable.
Method 1: Blur or pixelate before the screenshot leaves the page#
Most screenshots that need redacting are of web pages — a ticket, a customer's account panel, a broken admin screen. For those, you can mosaic the sensitive parts in the same motion as the capture, with Screenshot Annotation (free, Chrome and Edge, built by this site):
- Press
Alt+C(drag a region) orAlt+A(visible area) on the page. On macOS,Cmd+Shift+XandCmd+Shift+S. - Hover before confirming to snap to an element — useful when the sensitive area is a specific table column or card: the capture frame aligns to its actual bounds instead of your hand-dragged rectangle.
- Drag a mosaic over each sensitive area. The pixelation runs on the captured bitmap, in the editor, before anything is exported — drag over the emails, the prices, the customer names.
- Or drop a solid bar / dim the rest. For truly secret strings (keys, tokens), a black bar is stronger than any mosaic; the spotlight mode dims everything except the area you want readers to look at, which also hides surrounding context.
- Copy or download. The export is flattened — the mosaic is baked into the final pixels, so there is no layer to peel off later.

From the extension's demo: a capture snapped to page elements, PII fields numbered for the ticket, the API key pixelated.
Best for:
- Web UI screenshots carrying customer or account data into tickets, chat or docs.
- Cases where you'd otherwise paste the raw screenshot first and "redact it later" — which never happens.
Stated limits: it captures web pages only (not desktop apps), works in Chromium browsers, and the flattened export means you cannot reopen the markup to edit it — mosaic is forever, by design. The full annotation toolset is covered in How to take annotated screenshots in Chrome.
Method 2: On Windows, cover it — the built-in tools don't blur#
If you search for a blur tool in Windows' own screenshot tooling, you'll find a feature request on Microsoft's forums — Snipping Tool's markup offers pen, highlighter and crop, and no blur. Paint has no selective blur either. So the honest built-in answer on Windows is not blur at all — it's an opaque rectangle:
- Capture with
Win+Shift+Sand the screenshot lands on your clipboard. - Paste into Paint (still shipped with Windows 11).
- Draw a filled rectangle over the sensitive area — shape tool, fill and outline both solid black, drag to cover. A solid fill replaces the pixels underneath; there is nothing to un-darken.
- Copy the result back out and paste where it was going.
Best for:
- One-off screenshots of desktop apps, file Explorer, or anything outside the browser.
- Anyone who wants zero new installs — it's ugly, but it is true redaction.
💡 Pro tip: resist the highlighter in Snipping Tool's markup. It's the same semi-transparent-stroke trap as the phone marker — fine for pointing attention, wrong for hiding data.
Method 3: On macOS, use an opaque shape in Preview#
Preview's markup toolbar has shapes, text and sketch strokes — no selective blur for images. (Its dedicated Redact tool exists for PDFs only.) The destruction route on a Mac:
- Open the screenshot in Preview (double-click it, by default).
- Show the Markup toolbar (the pen-tip icon, or
Cmd+Shift+A). - Add a shape — the square — and set its fill to black, border to none or black.
- Drag it over the sensitive area. Because the shape's fill is opaque,
the covered pixels are replaced when you export (
Cmd+Shift+S, or just save). - Prefer a crop when the layout allows it:
Cmd+Kand cut the sensitive region out of the frame entirely.
Best for:
- One-off Mac screenshots, especially ones already destined for a ticket or email attachment.
The sketch and pen strokes in the same toolbar are the trap here — they're freehand ink layers, the exact thing the iOS marker warning below is about.
Method 4: On iPhone and Android, skip the marker#
The marker is the tool everyone reaches for on a phone screenshot — it's the first thing in Markup's drawer. It's also the documented failure case: the 9to5Mac PSA walks through how a black marker stroke over text can be re-revealed with a few taps in the photo editor, because the stroke is semi-transparent. The same applies to the highlighter. What to do instead:
- Open the screenshot in Photos and tap Edit → Markup.
- Tap the plus button and add a shape — the square or rectangle.
- Set its fill to solid black (tap the shape, the fill picker, the solid swatch) and drag it over the sensitive text. Opaque fill replaces what's underneath when the image is saved.
- Or crop instead: Edit → Crop, and cut the region out of the frame entirely. For a phone screenshot with one message to hide, cropping is usually cleaner than covering.
Best for:
- Phone screenshots — receipts, chat threads, settings pages — headed to anyone outside the conversation.
Method 5: Online blur tools — know what you're uploading#
Search "blur part of a screenshot" and the top results are online tools: drag, drop, brush over the area, download. They rank because they're genuinely easy. They also work by uploading the image you're trying to protect to a web server — you are handing the unredacted original, with its customer data or credentials, to a site you know nothing about, then trusting its retention policy.
A reasonable split:
- Mildly personal images (a face in a photo for a public post, a neighbor's license plate) — an online tool is a defensible convenience, and some do run their processing client-side in your browser. Check whether the tool still works with your network disconnected: if it does, it's local; if it errors, your image traveled.
- Anything with customer data, account details, keys or tokens — redact locally. Method 1 for web pages, Methods 2–4 for everything else. The convenience isn't worth the third party.
Best for:
- Images that are personal but not sensitive — and for which installing anything would take longer than the task.
Which method fits which situation#
| Method | True redaction | Effort | Works on | The catch |
|---|---|---|---|---|
| In-browser mosaic at capture (Method 1) | Yes — flattened into the export | Lowest | Web pages, Chrome/Edge | Web pages only; markup not re-editable |
| Opaque rectangle in Paint (Method 2) | Yes — replaces pixels | Low | Any Windows screenshot | It's a bar, not a blur — visually blunt |
| Opaque shape in Preview (Method 3) | Yes — replaces pixels | Low | Any Mac screenshot | Shape must be set to solid fill |
| Shape or crop in Markup (Method 4) | Yes — solid fill or removed | Low | Phone screenshots | The default marker tool is the trap |
| Online blur tool (Method 5) | Usually — but you uploaded the original | Lowest | Anything | The unredacted image leaves your machine first |
And by scenario:
| Situation | Reach for | Why |
|---|---|---|
| Web screenshot with customer data going into a ticket | Method 1 | Redact at capture; nothing to peel off |
| Screenshot of a desktop app on Windows | Method 2 | No built-in blur exists; the bar is real redaction |
| Screenshot on a Mac | Method 3 or crop (Cmd+K) | Same reasoning, Preview's shape tool |
| Phone screenshot | Method 4 (shape or crop) | Never the marker |
| Hiding a face in a public post, nothing sensitive | Method 5 | Convenience is fine when the stakes aren't credentials |
| Anything containing passwords, keys or tokens | Solid bar or crop | No blur, no fine pixelation, ever |
What to redact before you share#
The blur step is the second-to-last step. The last step is the scan — the things that ride along in screenshots more often than anyone notices:
- Email addresses — yours, your colleagues', the customer's.
- Names and usernames in comments, chat panels, account headers.
- API keys, tokens, and URLs with session parameters — a URL containing
?token=…is a credential in plain sight. - Prices and revenue figures — screenshots of dashboards leak business data sideways.
- Phone numbers and addresses, including in notification toasts.
- Other people's faces when the image is going somewhere public.
💡 Pro tip: if you might need the unredacted version later, keep a private copy on your machine before you export the redacted one. Redaction is one-way — that's its whole value — so don't rely on the chat history or the ticket to keep the original for you.
For the wider discipline of what makes a screenshot useful (and safe) in a bug ticket — framing, context, what to attach as text instead — see how to take a screenshot for a bug report developers can act on.
FAQ#
Can blurred or pixelated text be recovered?#
Yes, in documented cases. The Depix tool recovers text pixelated with the common linear box filter — demonstrated on monospaced passwords — and semi-transparent marker strokes can be re-revealed with brightness adjustments. Blur should be treated as a visual convenience, not a security measure: for secrets, use a solid bar, a coarse mosaic, or a crop.
Is the iPhone or Android marker tool safe for redacting?#
No. The marker stroke is semi-transparent by design, and text covered with it can be made readable again with basic editing adjustments — the exact failure 9to5Mac documented. Use a solid filled shape from Markup's plus menu, or crop the region out.
Is a blur safer than a black bar?#
The reverse. A black bar replaces the pixels — the information is gone. Blur and fine pixelation reorganize the information but keep it in the file, and research tools keep getting better at undoing them. Bar or crop for secrets; coarse mosaic for casual privacy; light blur never for anything sensitive.
Can I blur part of a picture in Word, Google Docs, or PowerPoint?#
None of them blur a region of a placed image — they can soften an entire picture with an artistic effect, which is a different thing. Redact the image before inserting it, using one of the methods above, and the document will carry the already-safe version everywhere it gets forwarded.
Do online blur tools upload my screenshot?#
Most of them do — the file goes to a server, gets processed, and comes back. Some run entirely client-side in your browser; the quick test is whether the tool still works with your network disconnected. For images containing customer data or credentials, use a local method instead.
Does Screenshot Annotation upload my screenshots?#
No. Capture, mosaic and export all happen locally in your browser; the finished image goes to your clipboard or your downloads, nowhere else. What it does and doesn't support is listed plainly on its product page — including the parts it can't do, like capturing desktop apps or full-page scrolls.
Editorial disclosure
We are the developers of the extensions documented on this site. This page describes the shipped version as we built and tested it — including its limitations — and was fact-checked on . Claims are based on our own testing of released builds; where third-party tools are named, their capabilities are described as documented by their publishers.
Related articles

How to Take a Screenshot for a Bug Report Developers Can Act On
How to take a bug report screenshot developers can act on: what to include, three ways to capture it, and redacting sensitive details before you attach.
9 min read · by Virgil Dean

How to Take Annotated Screenshots in Chrome
Capture a page, snap the frame to the exact element, and annotate with arrows, boxes, zoom callouts and mosaic — without leaving the browser tab.
4 min read · by Virgil Dean

PNG vs JPG vs WebP: Which Format for Screenshots?
Why screenshots should usually be PNG, when JPG is acceptable, what WebP wins at, and how each format handles UI elements, text and gradients.
3 min read · by Virgil Dean
Extensions mentioned in this article
Every tool we build is documented with its limits stated plainly.