EXIF metadata remover

What is actually hidden in a photo

Image formats are containers. A JPEG is a chain of marker segments, a PNG is a chain of chunks, a WebP is a chain of RIFF chunks — and in every case the compressed pixels occupy only some of them. The rest are free to hold whatever the camera, phone or editor felt like recording. In a JPEG that is usually an APP1 segment holding Exif, sometimes a second APP1 holding Adobe XMP, an APP2 with the colour profile, and an APP13 with the IPTC caption and credit fields that photo agencies use. In a PNG it is tEXt, iTXt and zTXt chunks, an eXIf chunk and a tIME timestamp. In a WebP it is chunks literally named EXIF, XMP and ICCP.

The Exif block is the one worth reading. Inside it is a small TIFF file with its own byte order, its own directory structure and its own pointers, and it commonly records the make and model of the camera, the software that last touched the file, the date and time to the second, the orientation the sensor was held at, and — if location services were on — a GPS directory with latitude and longitude stored as three fractions each: degrees, minutes and seconds. This page converts those fractions into ordinary decimal coordinates and prints them, because seeing your own street appear under a holiday photo is more persuasive than any paragraph about metadata.

None of this is visible when you look at the image, and almost none of it is removed by the usual habits. Renaming a file keeps it. Copying it to a USB stick keeps it. Sending it through most chat apps strips some of it, but you do not get to see which, and a file attached to an email or uploaded to a forum usually arrives complete. The only reliable answer is to look at the container yourself.

Why removal here is lossless

There are two ways to strip metadata. The common one is to decode the image into pixels, draw it onto a canvas and encode it again — easy to write, and it does drop everything, because the new file is built from nothing but pixels. It also re-compresses the image. For a JPEG that means a second generation of lossy compression on top of the first, so fine detail softens and edges pick up artefacts, and the file often changes size in the wrong direction.

This page does the other thing. It walks the container, decides which blocks go, and copies the survivors byte for byte into a new file. In a JPEG, every segment before the start-of-scan marker is either copied verbatim or dropped, and everything from the start-of-scan onward — the entire compressed image — is copied in one piece without being looked at. In a PNG, each surviving chunk travels with its own CRC, so no checksum needs recalculating and none can silently disagree. In a WebP the surviving chunks are copied and only the RIFF length field and the VP8X feature flags are rewritten to match the shorter file.

The result is that the image data in the cleaned file is identical to the image data in the original, bit for bit. Open both in an editor and no pixel has moved. The file is smaller purely because the metadata is gone, which on a phone photo is typically a few tens of kilobytes, and on a file that carries an embedded thumbnail can be a great deal more.

Questions people ask

Does this reduce image quality at all?

The compressed pixel data and compression quality do not change. The picture can still look rotated when its Exif Orientation tag is removed, and colours can shift if you choose to remove an ICC profile. This is the difference between a metadata stripper and a converter: a converter rebuilds the image from pixels and gets a new, slightly worse file, while a stripper edits the container around the image and leaves the compressed image data alone. The other changed bytes are the metadata you asked to remove and the length fields that describe the file's own structure.

Why did my cleaned photo come out rotated?

Because the orientation tag lived inside the Exif block that was just removed. Phone cameras often store the image the way the sensor read it and attach a note saying which way up it should be shown. Remove the note and viewers fall back to showing the raw pixels, which can be ninety or a hundred and eighty degrees out. Rotating the pixels to compensate would mean decoding and re-encoding the image, which is the one thing a lossless strip must not do, so this page does not attempt it. If a cleaned photo looks wrong, rotate it once in any photo application and save it; from then on it is correct without needing a tag.

Should I keep the colour profile?

Usually yes, and the box is ticked by default because it preserves colour. An ICC profile describes which physical colours the numbers in the file refer to. Photos from recent phones are often in Display P3 rather than sRGB, and if you delete the profile a viewer will assume sRGB and show noticeably duller, shifted colours — reds in particular. A custom profile can also contain device names, dates, descriptions or copyright text. Keep it for colour accuracy; untick it when removing those details matters more.

Does removing metadata make a photo anonymous?

It removes one channel of identification, not all of them. The pixels themselves still show whatever they show, including faces, house numbers, reflections and screens. Sensor noise patterns can in principle link photos to a particular camera, and the way an image was compressed can hint at the software that produced it. Beyond the file, the platform you upload to records your account, your address and the time. Stripping Exif is a sensible, cheap precaution against the specific and very common problem of publishing your home coordinates by accident; treat it as that rather than as anonymity.

Which files can this page read?

JPEG, PNG and WebP. Those three cover almost everything that comes off a phone, a screenshot tool or a website. Camera raw files such as CR2, NEF and DNG are not supported, and neither is HEIC, the format iPhones use by default when High Efficiency is selected — an HEIC file is an ISO base media container and needs a different parser. If a file is not one of the three, the page says so rather than handing back something it did not really clean.

Why does the file get smaller when nothing was compressed?

Because metadata takes real space. A typical Exif block is a few kilobytes, but many cameras also embed a full JPEG thumbnail inside it, and XMP blocks written by editing software can run to tens of kilobytes of XML. Removing those does not touch the image and yet often takes a noticeable slice off the file. If you want the image itself to shrink, that is a separate job — use the compressor pages linked below, which do re-encode deliberately.