Share the image. Keep the details private.
Cover sensitive information before you share a screenshot, document, or photo.
● Local processing — your image and detected text are not uploaded by this tool.
- 1 Choose image
- 2 Select & review
- 3 Redact & download
Choose an image
JPEG, PNG, or static WebP · up to 20 MB, 12 megapixels, and 4096 pixels per side.
One image at a time. No account needed.
Choose an image to begin. Manual selection works even without text detection.
Click boxes to select multiple fields. Draw over anything missed. Drag the corner of an active box to resize it.
Review your redacted image
Check the whole image before sharing. The download contains the covered pixels, not editable boxes.
The suffix _aixsafe.png is added automatically. Filenames may contain personal information; rename yours if needed.
Exported as a new PNG without copying original file metadata. Transparent areas become white.
What to expect
Does my image leave my device?
This editor processes images locally and does not send images, filenames, or detected text to a server. It downloads the site and OCR software from AIxSafe. Hosting providers receive normal website request information. The editor has no analytics or session replay.
Will it find every piece of personal information?
No. English OCR suggests emails, phone numbers, and some labelled identifiers. It may miss or misread text. Names, addresses, handwriting, faces, signatures, and QR codes need your review and manual selection. No suggestions does not mean the image is safe to share.
How does redaction work?
Selected areas are covered with opaque pixels in a new flattened PNG. Optional XXXX text is drawn on top. Blur and mosaic are not used. The original file on your device is unchanged.
Why is my image unsupported or slow?
PDF, SVG, GIF, HEIC, and animated images are not supported. Large images need more device memory. Use a smaller still image if processing fails. Text detection can be cancelled; manual redaction remains available.