Can blurred or pixelated areas in a saved image be undone?
Pixelated, blurred, and blacked-out regions overwrite the canvas pixels directly before export, so the saved file contains no reversible layers or hidden original pixels. However, light blurs over short, predictable strings like passwords or account numbers can sometimes be guessed, which is why Black box is safer for sensitive text.
According to NIST SP 800-88 Rev. 1 Guidelines for Media Sanitization, effective sanitization must leave no recoverable data signal rather than relying on weak masking. If you need to annotate a censored screenshot for a bug report, you can label redacted screenshots with callout notes after saving.
Permanent pixel replacement versus reversible layers
Desktop image editors often keep redactions on separate layers that can be toggled off if saved in a project format. This tool flattens every Pixelate, Blur, and Black box region into a single raster grid, removing all layer history.
In Blur mode, the canvas explicitly runs clearRect on each box before drawing the two-pass downsampled blur. That step prevents sharp edges on semi-transparent PNG screenshots from bleeding through the blurred overlay.
Because every pixel inside the rectangle is overwritten in memory before encoding, there is no undo data, alpha mask, or hidden layer stored in the downloaded JPG, PNG, or WebP file.
Why Black box is safer than Blur for short text
Standard fonts have predictable character heights, kerning, and word lengths. Even when averaged into mosaic blocks or smoothed by downsampling, short numbers or API keys can leave statistical clues for brute-force guessing attacks.
Choosing Black box replaces every pixel inside the rectangle with solid #000000, reducing entropy to zero so nothing inside the box can be reconstructed.
Choosing between Pixelate, Blur, and Black box modes
Pick Pixelate to blur faces and number plates in photos privately with an unmistakable mosaic look, Blur for softer background obscuring, and Black box for passwords, emails, or financial IDs. Strength (1 to 12, default 5) tunes Pixelate and Blur, while Black box is always 100% opaque.
| Effect mode | How it alters pixels | Adjustable strength | Best use case |
|---|---|---|---|
| Pixelate | Averages RGBA values into square blocks | 1 to 12 (default 5) | Faces, license plates, and street signs |
| Blur | Two-pass downsample and smooth upsample | 1 to 12 (default 5) | Soft background obscuring and whole photos |
| Black box | Solid #000000 opaque rectangle fill |
None (always 100% solid) | Passwords, API keys, and account numbers |
| Whole picture | Applies chosen effect to full canvas | Uses active mode setting | Teaser previews or full-frame obscuring |
Under the MDN CanvasRenderingContext2D.getImageData() Reference, Pixelate reads raw RGBA buffers inside a Web Worker to pixelate sensitive information in screenshots locally. For multi-page documents with selectable text layers rather than screenshots, use a browser-based PDF redaction tool instead.
Drawing, resizing, and nudging boxes on the stage
You can draw as many redaction boxes as you need directly over the 720 px preview stage, or click + Add area to drop a centered box that you can drag into position. Clicking Clear areas removes all boxes at once if you want to start over.
Using Whole picture and Strength from 1 to 12
Ticking Whole picture hides individual boxes and applies the selected Effect across the entire image dimensions. This is useful when creating blurred background cards or obscuring an entire document preview.
The Strength slider scales block size dynamically based on the image’s shorter side divided by 160. Because it scales proportionally with image dimensions, a Strength of 5 looks consistent on both a small screenshot and a 4K photo.
Drawn rectangles are clamped to a minimum side of 8 px (or 3% of the image’s shorter edge) so tiny accidental clicks do not leave invisible 1-pixel boxes behind on the stage.
Keyboard shortcuts for precise box placement
Every drawn box is keyboard focusable for accurate alignment without fighting mouse drift. Press Arrow keys to nudge a box by 2.5% of the image’s shorter side, or hold Alt with Arrow keys for 0.5% micro-adjustments.
Hold Shift plus Arrow keys to resize the box’s width or height, and press Delete or Backspace to remove the focused box. If you are preparing scanned paperwork, you can also convert redacted document scans into high-contrast monochrome to shrink file size and improve legibility.
Local browser processing and automatic metadata removal
Every pixel transformation runs inside your browser tab, so sensitive screenshots containing banking details, medical records, or private messages never touch a remote server. Images larger than 4096 px on a side or 16,777,216 total pixels are clamped proportionally to fit browser canvas memory limits.
Keeping confidential screenshots off remote servers
Uploading unredacted photos to cloud editors exposes the original file before you even draw a mask. Here, Pixelate runs in an isolated Web Worker on your device, and Blur and Black box render directly on your local 2D canvas.
Tip: Before sharing a photo of a document or vehicle, check both the visible corners and reflected surfaces, and use Black box over any alphanumeric code or QR pattern.
Automatic EXIF and GPS coordinate stripping on export
As explained in the MDN HTMLCanvasElement.toBlob() Reference, exporting the canvas at 0.92 quality produces a brand-new raster blob (-blurred or -redacted) with zero original EXIF or GPS tags. To strip location data from photos you do not need to blur, you can clear hidden GPS coordinates from untouched photos directly.