How it works

Short version: your browser reads the photo, a small program inside this page removes the hidden data, and you download a clean copy. This site has no way to receive a file.

What happens to a photo

You choose a photo. The browser lets this page read it. Nothing has left your device.
The page works out the file type from its contents, lists the hidden data it finds, and shows what each item could reveal.
When you press Remove, it writes a new file that keeps the picture data exactly as it was and leaves the hidden data out. It reads the new file back to confirm nothing is left, and compares a checksum of the picture data before and after.
You download the clean copy. The original is not touched.

What is removed

JPEG: Exif (including GPS and maker notes), XMP, IPTC, comments, embedded thumbnails, vendor application data, and anything stored after the picture, such as the video in a motion photo.
PNG: text notes, time stamps, Exif, XMP, unknown extra chunks, and anything after the end of the image.
WebP and GIF: Exif, XMP, comments, extra application data.
HEIC and AVIF: the Exif and XMP items are overwritten in place with empty ones of the same length, and extra boxes such as a motion-photo video are cut off or blanked. Nothing else in the file moves, so the picture is untouched. The file does not get smaller.

What is kept: the picture, and by default the rotation hint and the colour profile. You can turn both off.

Check it yourself

Watch the network. Open your browser's developer tools (F12), choose the Network tab, add a photo and press Remove. You will see requests for this site's own program files at most, and none that carry your photo.
Go offline. After the page has loaded, switch off Wi-Fi. The tool still works.
Use a second reader. If you have ExifTool, run exiftool -a -G1 -s your-photo-clean.jpg. A clean copy lists file basics and, if you kept them, the rotation and colour profile. No GPS, serial number, date or name.
Read the meter. The panel at the bottom of every page lists what the page loaded and flags any request that carries data or goes to another site. It sees only requests made by this page's own scripts, so treat it as a convenience and the first three checks as the proof.
Read the policy. The page's Content Security Policy tells the browser to refuse connections to anywhere but this site. The browser enforces it, not us.

What we do see

Cloudflare delivers this site and, like any web host, records ordinary access logs: the visitor's address, the pages and files requested, and the time. We add no analytics, cookies or tracking scripts. The names and contents of your photos are never part of those requests.

What it cannot do

It cannot remove what is in the picture: a street sign, a face, a reflection, a house number.
It cannot change the file's name. If the name says "Maria-birthday-Lisbon.jpg", rename it before you share it.
It does not touch video files.
It lists and removes the places hidden data is known to live. A format that stores something somewhere we do not know about would not be seen. That is why the result is read back and why we suggest a second reader.
Social sites and messaging apps may remove some of this from the copies they show, and may keep it in the originals they store. You cannot see which. A clean copy removes the question.
A phone's own "share without location" switch removes location only. This tool removes the rest as well.

Limits

Up to 500 MB per file. Very large batches may run out of memory on a phone; if that happens you will see a message rather than a crash. Always keep your originals.

What is inside

The readers and writers for each file type are our own code, written from the public format specifications, and use no third-party software in your browser. The engine build number in the page header changes whenever that code changes, and engine.json lists every file with its checksum. Open source notices are in licenses.txt.

This page was built from engine ….