Blur faces online — automatic single and batch face blurring
Hide faces before sharing a photo. Let the detector mark a first set of faces, then check the picture yourself, add any it missed and choose blur, pixelation or a solid cover.
Runs in your browser — your input is not uploaded to ToolsNow.
Proportional to each face
JPEG and WebP only
Your files are processed locally in your browser and are not uploaded to ToolsNow.
Site files and analytics use separate network connections. Ads are off. Privacy & Cookies
Your photo is processed in your browser, without uploading it to ToolsNow. Detection runs in your browser after the detector files load. It finds faces, not identities.Automatic detection may miss small, obscured, angled or partially visible faces. Always inspect the completed image before publishing or sharing it. Blur and pixelation can leave people recognisable. Use solid cover where needed and check the downloaded image, including uncovered details. Face covering does not guarantee anonymity.
Blur, pixelate or cover?
| Effect | What to check |
|---|---|
| Blur | Softens detail using the original image. A face can still be recognisable; a high strength setting is not proof that someone is hidden. |
| Pixelate | Makes a blocky mosaic from image samples. It keeps information from the source, so inspect the downloaded file rather than judging by the preview alone. |
| Solid cover | Paints an opaque colour over the selected shape. Choose this when facial detail needs to be removed, and make sure the shape covers it fully. |
An Ellipse leaves the corners of its bounding box outside the effect. Rectangle covers the whole box; Rounded rectangle leaves curved corners. Check the actual coverage around the forehead, ears and chin, not just the outline.
Give a group photo a second pass
If the tool marks three people in the foreground, scan the background and edges separately. Look for a face between shoulders, a profile at the edge of the frame or someone reflected in a window. Add boxes for missed faces, then check the exported image at full size.
Covering faces does not make the scene anonymous. A name badge, tattoo, uniform, location or caption may still identify someone. Use image redaction for other details that need covering.
Reviewing a batch
Batch applies the chosen settings automatically; a completed row is not a human review. Its Review button opens the original in Single image mode. If you make corrections there, use Download blurred image to save them separately: those edits do not update the batch ZIP. Check every file you intend to share, including rows with no faces found.
How to use
- Open a photo. Choose Single image to edit a JPEG, PNG or WebP. Wait for detection to finish; the first image also loads the detector.
- Set the treatment. Choose Effect, Region shape and Region padding before making manual corrections. These settings apply across the image.
- Check the whole frame. Look beyond the outlined faces. Drag on an empty area to add a missed region; move an existing box or use Bigger and Smaller in the region list.
- Check which regions are enabled. A region labelled Skipped receives no effect. Remove false detections, but inspect faces without a “needs review” label too.
- Download and inspect the file. Choose Output format, then Download blurred image. Open that download at full size and check the faces and edges before sharing it.
Questions
What should I do if it says “No faces detected”?
Check the photo yourself and draw regions over faces that need covering. Small background faces, profiles, partly hidden faces and poor lighting can defeat detection. An empty result means the detector found no boxes, not that the photo contains no people.
What does the confidence percentage mean?
It is the detector’s score for a face-shaped region, not a measure of how well someone is hidden. Detections below 50% are marked “needs review”. Higher-scoring boxes still need checking for coverage. The tool locates faces; it does not identify the people in them.
Can I change the padding after drawing my own boxes?
Set Region padding first if possible. In the current editor, changing it can restart detection and replace manual corrections, so recheck the region list afterwards. Detect again also replaces the current regions. Use Bigger or Smaller for an individual box once you are making corrections.
What are the file and batch limits?
Batch accepts up to 100 JPEG, PNG or WebP files, with a 25 MB limit per file. Both modes reject images above 40 megapixels. Batch processes files one at a time; Cancel stops further processing and completed downloads remain available. A large image can still exceed the memory available on your device.
Which output format should I choose?
Same as input keeps the format for JPEG, PNG and WebP files. You can also choose one of those formats explicitly. PNG and WebP can retain transparency; JPEG fills transparent areas with white. Quality affects JPEG and WebP only. The download is a new image, not an edit to the original file.
Does the download keep the original photo’s metadata?
The image is drawn onto a canvas and encoded afresh, so the original EXIF metadata, including any GPS tags, is not copied into the download. Visible clues such as street signs remain. The filename also keeps the original name’s stem, with “-faces-blurred” added, so check that before sharing.
Published by ToolsNow. How tools and sources are checked · Report a problem