Redact images and screenshots before you share them
Screenshots travel everywhere — into support tickets, group chats, bug reports, social posts and slides — and they often carry more than you meant to share: an email address in a header, a phone number in a chat, the last digits of a card on a payment page, an Aadhaar or PAN number on an ID photo. This redact image tool covers that information properly before the picture leaves your hands.
You can draw boxes by hand, or let it find the text for you: it reads the image with OCR on your own device, then automatically boxes email addresses, phone numbers, card numbers, Aadhaar and PAN numbers, IBANs, US Social Security numbers, and any words or patterns you type. Every box can be checked, moved, resized or removed before you download.
Choose a solid box (the safest), pixelation or blur for each area. The download is a flattened image — the redactions are part of the pixels, with no layers, no hidden original and no EXIF metadata — and nothing is ever uploaded.
How to use the Redact Image
- Drop a screenshot or photo onto the upload area, click to choose one, or paste a screenshot with Ctrl+V.
- Optional: tick the kinds of personal data to look for, type any names or words, choose the language and press Find and box text.
- Check every box on the image. Drag on empty space to add a box, drag a box to move it, drag its corner to resize it, and press × to remove it.
- Pick the effect — Solid box (safest), Pixelate or Blur — for the selected box or for all boxes, and a box colour.
- Choose a format and press Download redacted image. PNG keeps screenshots crisp.
Redact Image features
OCR “find text”
Tesseract OCR runs in your browser, so the tool can locate text in a picture and box the words that match, even inside a longer line — tolerating the stray spaces OCR sometimes adds to email addresses.
Personal-data presets
Emails, phone numbers, payment cards (Luhn-checked or in card-style groups), Aadhaar, PAN, IBAN and US SSN patterns, plus your own words or regular expressions.
Flattened output
Effects are burned into the pixels and the image is re-encoded, so there are no layers or hidden originals, and camera metadata such as GPS is dropped.
Three effects, per box
Solid box in any colour, pixelation or blur, chosen per box — for example blur on a face and solid black on a card number.
Precise editing
Move, resize, nudge with the arrow keys, delete, undo and clear, with mouse, pen or finger.
Private
The image and the recognised text stay on your device; only the OCR engine and language data are downloaded once.
Why blur and pixelation can be reversed, and solid boxes cannot
Blurring and pixelation are mathematical averages of the original pixels, so some information always survives. That is enough for an attacker who can guess what kind of text is underneath: if it is a card number in a known screenshot font, they can render every likely number, pixelate or blur it the same way, and see which one matches. Open-source tools such as Depix do exactly this for pixelated screenshots, and deblurring software can sharpen weak blurs. Larger blocks and stronger blur make it harder, not impossible.
A solid box replaces the pixels completely — there is nothing left to compare against. Use solid boxes for passwords, account, card and ID numbers, signatures and anything with legal or financial weight. Pixelation and blur are fine for faces in a crowd or a background you simply want to de-emphasise; for faces, the Face Blur tool can find them automatically.
What the automatic search looks for
| Preset | Example | Rule |
|---|---|---|
| Email addresses | name@example.com | Standard address format |
| Phone numbers | +91 98765 43210, (555) 123-4567 | 9–15 digits with separators, or a 10-digit Indian mobile |
| Card numbers | 4111 1111 1111 1111 | 13–19 digits passing the Luhn check, or grouped 4-4-4-4 / 4-6-5 |
| Aadhaar numbers | 2345 6789 0123 | 12 digits starting 2–9, grouped 4-4-4 or unbroken |
| PAN numbers | ABCPE1234F | Five letters, four digits, one letter |
| IBAN | GB82 WEST 1234 5698 7654 32 | Country code, two check digits, 11–30 more characters |
| US SSN | 123-45-6789 | Dashed format, invalid ranges excluded |
Limits of OCR — always check by eye
- OCR reads printed and on-screen text well, but small, blurry, stylised or handwritten text can be misread or missed entirely. Review the whole image before sharing.
- If a number is split across two lines, or OCR reads a 0 as an O, the pattern may not match — draw a box by hand.
- Text inside logos, QR codes and barcodes is not read; cover codes manually, since they can hold personal data too.
- Remember what is around the text: profile photos, names in tabs, notification banners, reflections. To remove whole regions, cropping with Crop Image is even safer.
- For PDF documents use Redact PDF, which also removes the hidden text layer.
Redact Image FAQ
How do I redact a screenshot?
Paste or drop the screenshot here, press “Find and box text” to box emails, phone and card numbers automatically, add any boxes you need by dragging, keep “Solid box” selected and download. The redaction is burned into the pixels.
Can blurred text be unblurred?
Sometimes. Weak blur and small pixel blocks over predictable text, such as numbers in a common font, can be reconstructed by trying candidates. Solid boxes cannot be reversed, so use them for anything sensitive.
Is the hidden text really gone from the file?
Yes. The download is a new, flattened image made from the edited pixels. There are no layers, no copy of the original and no EXIF metadata in it.
Does the automatic search find everything?
No tool can promise that. OCR can miss small or unusual text, and patterns can miss odd formats. Treat the automatic boxes as a head start and check the whole image yourself.
Can it redact Aadhaar and PAN numbers in a photo of my card?
Yes. Tick “Aadhaar numbers” and “PAN numbers” and press Find and box text. For photos taken at an angle, OCR may struggle — draw boxes by hand where needed.
Is my image uploaded?
No. OCR and redaction run in your browser. Only the Tesseract engine and language data are downloaded from a public CDN; the image never leaves your device.