File picking and downloads often fail inside apps. Open this page in your default browser (Chrome, Safari) or on a computer.
Blur & Pixelate Image
Before you post a screenshot, a marketplace listing or a group photo, hide faces, phone numbers, addresses and license plates with pixelation, blur or a solid box. The photo never leaves your browser, and the saved image carries no EXIF data such as GPS location.
Choose a photo, drag over what you want to hide, then press Create censored image.
How to blur or pixelate part of a photo
- Press Choose a photo. The photo appears in the editor below the options.
- Pick an effect. Pixelate or blur suits faces; for text such as phone numbers, addresses, card or account numbers, use Solid.
- Drag a box over each area, or switch to Brush and paint over it. Press Undo if you miss.
- For screenshots and labels with text, press Find personal details automatically. Matches appear as red boxes; keep the ones you want ticked and press Cover selected.
- Press Create censored image, check the result and download it. The pixel dimensions stay the same as the original.
Pixelate, blur or solid box: which is safe?
| Effect | Looks like | Best for | Watch out |
|---|---|---|---|
| Solid | A black (or chosen color) box | Phone numbers, addresses, names, card and account numbers, ID numbers | The original is gone for good; nothing to recover |
| Pixelate | Large color blocks | Faces, license plates, bystanders | Small blocks over text can be reversed by guessing the characters |
| Blur | A soft smear | Backgrounds, signs, anything you only want to tone down | Light blur keeps outlines and can be partly undone |
Pixelated text is not as safe as it looks: tools such as Depix rebuild it by matching the block pattern against known fonts. Cover personal text with a solid box, and when you pixelate, keep the strength on Strong. The original pixels of a covered area are not kept anywhere in the saved file.
How does automatic detection work?
The button runs Tesseract.js, an OCR engine, inside your browser. It reads the text in the photo and marks anything that looks like a phone number, email address, card or bank account number, US Social Security number or street address. The photo itself is never uploaded; only the engine and an English text model (about 5 MB) are downloaded once and then cached.
Treat the boxes as suggestions. Handwriting, tiny or blurry text, crumpled or tilted labels, barcodes and QR codes are not detected, and names are only found after a label such as "Name:". Always look over the photo afterwards and drag over anything that was missed. Detected text is covered with a solid box, so the original characters are not kept in the saved file.
What to hide before you share
- Marketplace listings: shipping labels (name, address, phone), receipts and warranty cards showing card numbers, mail or street views visible in photos of your home.
- Social media and blogs: faces of people who didn't agree to be posted, children's name tags and school uniforms, license plates, house numbers.
- Work screenshots: chat names and avatars, customer contact details, internal URLs, passwords or one-time codes on screen.
Photo files can also hide data you can't see, such as the GPS position where the picture was taken and the phone model. The image saved here is redrawn from scratch, so that EXIF data is removed too.
FAQ
Is my photo uploaded?
No. Opening, editing and saving all happen with the browser's canvas on your own device. The photo is never sent to or stored on a server.
The page scrolls when I try to select on my phone.
Dragging on the photo selects an area instead of scrolling; to scroll the page, drag outside the photo. For small text, rotating the phone to landscape or using a computer makes precise selection easier.
Does it find faces automatically?
No. Automatic detection only looks for personal details written as text. Drag over faces yourself, or turn on Apply to the whole photo to blur everything.
Does automatic detection send my photo anywhere?
No. The text recognition engine is downloaded and runs on your device, so the photo never leaves it. Only the engine files are downloaded; nothing about your photo is sent.
The saved file is larger than the original.
PNG files are saved as PNG and JPEGs as JPEG at quality 92. If the site you are posting to has a size limit, run the result through Compress Image.
Sources
- In-browser OCR engine: Tesseract.js (GitHub)
- Recovering pixelated text: Depix (GitHub)
- Photo location data and privacy: FTC, How To Protect Your Privacy on Apps
- Canvas image processing: MDN, Pixel manipulation with canvas
Found this useful? Share it
Only the page link is shared. Your inputs and files are never included.