IT Image Tools

License Plate Blur — Hide Number Plates in Photos

Automatically find and blur, pixelate or black out car license plates in photos, right in your browser. Toggle boxes, add missed plates and download — no uploads.

🔒 Runs entirely in your browser — nothing is uploaded

Drop a photo here or click to browse

Click a box to turn blurring on/off. Drag on the photo to cover a plate that was missed. Higher sensitivity shows weaker matches.

On first use a vehicle detection model is downloaded once from Google's public model storage (the detection engine comes from this site). Your photo is processed on this device and never uploaded.

Advertisement

Hide license plates before you post a photo

Car listings, street photography, dashcam stills, accident reports and holiday snaps often show readable number plates. A plate can be linked to a vehicle owner, which is why many platforms, insurers and privacy rules such as the GDPR expect it to be hidden before an image is published. This license plate blur tool finds the plates for you and covers them in one step. Every match gets a box you can switch on or off, so you can keep your own car's plate visible in a sale ad while hiding the cars around it, and you can drag extra boxes over any plate the detector missed.

Detection runs in two stages, both inside your browser. First, Google's MediaPipe object detector finds cars, trucks, buses and motorcycles. Then each vehicle is scanned for the typical pattern of a plate: a short row of high-contrast characters with regular dark and light strokes in the lower part of the vehicle. The vehicle model is downloaded once from Google's public model storage and the engine is loaded from this site; your browser can cache both. The photo itself never leaves your device.

Blur, pixelate or black box?

The black box is the safest choice because it removes the characters completely. Blur and pixelate look less intrusive and work by shrinking the plate to only a few samples before enlarging it again, which permanently throws away the detail needed to read it. Keep the strength at 6 or higher: light blurs on large, sharp plates can sometimes still be deciphered. The boxes are slightly larger than the characters so the plate frame and the country band are covered as well.

Getting reliable results

The default mode only looks at plates on detected vehicles, which avoids blurring shop signs and street names. If a car is only partly visible or very small, switch to Whole image to scan the entire photo, then turn off any boxes on unrelated text. The sensitivity slider shows or hides weaker matches instantly without detecting again. Plates seen at steep angles, in shadow, covered in dirt or only a few pixels tall are hard to find automatically, so always check the preview before you share. The saved file contains no boxes and no original EXIF metadata such as GPS location.

How to use

  1. Add a photoDrop a JPG, PNG or WebP photo; vehicles and their plates are detected automatically.
  2. Review the boxesClick a box to include or exclude it, or drag to cover a plate that was missed.
  3. Pick the effectChoose blur, pixelate or a black box and set the strength.
  4. DownloadSave the anonymised image in its original format.

Frequently asked questions

Is my photo uploaded to find the plates?
No. Your photo is processed entirely on your device. Only the vehicle detection model is downloaded, once, from Google's public MediaPipe model storage; the image itself is never sent anywhere.
What if a plate is not detected?
Raise the sensitivity slider to show weaker matches, or switch Detection to Whole image. You can always drag a rectangle over any plate that was missed to blur it manually.
Why did it blur a sign or some text?
Plates are found by looking for rows of high-contrast characters, so other text can look similar. Click the box to switch it off, lower the sensitivity, or keep Detection on vehicles so only text on cars is considered.
Which effect hides a plate best?
The black box removes the number completely. Blur and pixelate look more natural; use a strength of 6 or more so the characters cannot be read or reconstructed.
Does the downloaded image keep EXIF data?
No. The image is re-drawn on a canvas and saved as a new file, so camera metadata such as GPS location is not copied.
Advertisement