Hide sensitive details
Privacy Blur
Blur faces, licence plates, screens, or any area before sharing a photo. Drag the blur regions directly on the preview — no uploads, no accounts.
Everything runs in your browser. Your photo never leaves your device.
① Drop your photo② Drag blur regions on the preview③ Adjust strength & download
Preview
Preview appears here
Drop a photo — then drag blur regions on the preview to position them.
Add blur region
Blur strength
18pxExport
Quality
92%Drop a photo above to enable download.
Next steps
Edit photo
RecommendedAdjust brightness, apply filters, crop, and edit — right in your browser.
Add watermark
RecommendedProtect your images — add text or logo watermarks with full control.
Beautify screenshot
Turn plain screenshots into polished visuals with frames and gradients.
Compress image
Shrink images by up to 80% without visible quality loss.
Runs entirely in your browser. No uploads. Your files stay private.
Selectively Blur Faces, Plates, And Sensitive Data
Privacy Blur applies a localised Gaussian-style blur to chosen regions of an image using the Canvas 2D API's filter property. Each region is defined as a rectangle or ellipse path on the source image, the path is set as a clipping region with Canvas's clip method, and the blur is applied only inside the clip — so the rest of the image stays pixel-identical to the original.
Regions are stored as relative coordinates (0 to 1 in each axis), not pixel positions. This means the blur regions scale perfectly when the source image is rendered at its native resolution for export. A region drawn on a 1920-pixel preview ends up in exactly the right place on a 24-megapixel original because the coordinates are resolution-independent.
Blur strength is configurable per region. The implementation maps the strength slider to a Canvas filter blur radius in pixels — values around 8-16 pixels are the right setting for hiding faces in average photos; 30-50 pixels for completely obliterating text in screenshots. Higher radii are slower because Canvas blur is implemented as a separable convolution that scales linearly with radius.
Multiple regions can stack on a single image. Each region is independently positioned, sized, and tuned, so you can use a tight ellipse for one face and a wider rectangle for a licence plate in the same photo. Up to eight regions are supported in the UI to keep the controls manageable.
Important: blurring is not the same as redacting. A determined attacker with the right tools can sometimes partially reverse a moderate blur using deconvolution, especially on text. For genuinely sensitive content (passwords, API keys, full home addresses), use a solid-colour rectangle rather than a blur — Photo Editor lets you draw shape overlays for that.
The blur is baked into the exported file. The Canvas re-encode happens at the original resolution with toBlob, producing PNG or JPEG output that has no recoverable original underneath. Keep a copy of the source elsewhere if you might need to re-edit later, because the export does not preserve the unblurred regions.
Everything happens locally. The image is loaded as an object URL, all blur regions are drawn into a Canvas in your tab, and the exported Blob is downloaded directly. There is no upload step, which is critical for the kind of content this tool is built for — photos with bystanders, documents with personal data, screenshots with credentials.
Common Use Cases
01
Faces of bystanders in photos
Blur faces of people who did not consent to appearing in a public post before sharing on social media or a blog.
02
Children's faces in family photos
Hide kids' identities in photos shared on community platforms while keeping the rest of the scene visible.
03
Licence plates
Blur licence plates in driveway, parking lot, or street photos where vehicle ownership could be traced.
04
Document redaction
Cover account numbers, addresses, or signatures on scans of bank statements, leases, or government forms.
Frequently Asked Questions
The region path is set as a Canvas clip before the blur filter is applied, so only pixels inside the clip are touched. The rest of the image is copied straight from the source bitmap and remains identical to the original.
Yes. Up to eight regions of any mix of rectangles and ellipses can be layered on a single photo. Each region has its own position, size, and blur strength.
Blur strength maps to a Canvas filter radius in pixels. Around 8-16 is enough to hide a face in a typical photo; 30-50 obliterates text in a screenshot. Larger radii are noticeably slower because Canvas blur scales linearly with radius.
Practically, no — the export bakes the blur into the pixel data. Theoretically, deconvolution can sometimes partially recover a moderate blur, especially on text. For genuinely sensitive content, use a solid-colour rectangle in Photo Editor instead.
No. Regions are stored as relative coordinates (0 to 1 in each axis), so they scale perfectly with the image. The same regions apply whether you export at 1080 px or at the original 24 MP.
Yes. Tap to move the selected region and use the sliders to resize it on touch screens. The blur preview updates in real time and the export works the same way as on desktop.
No. Everything runs locally in your browser using the Canvas API. The image, the blur regions, and the export are all kept inside your tab.
PNG (lossless) and JPEG (smaller). PNG is the safer default because it does not introduce additional compression artefacts on top of the blurred regions.
Some mobile GPUs implement Canvas filter blur with a smaller kernel than desktop browsers. Increase the blur radius slightly on phones if the result looks insufficient, or export and check the result on a desktop.
No — the export is the new ground truth. Always keep a copy of the original file before applying blur if you might want to reposition or remove regions later.
Advertisement