EXIF is not one tag and not one blob: it is a set of directories, each a table of tag ids and typed values, written into the same JPEG segment as an embedded thumbnail. This page is the technical companion to the photo tool — a reference to what the tags are, which ones identify you, and what actually happens when a document editor, a messaging app or a social platform rewrites the file.
Engine
File API
Runs in
Web Worker
Needs account
No
Tag ids are given in hex. The three directories below are the ones that matter for privacy; a fourth, the maker note, is a vendor blob this tool removes but does not decode.
A TIFF-structured EXIF block starts with a byte-order mark, an offset to the first directory, and then a table of entries: tag id, type, count, and either an inline value or an offset to one. There are separate directories for the main image (IFD0), the capture details (Exif IFD), GPS, and interoperability. “Has GPS” is really “IFD0 links to a GPS IFD”.
IFD1 holds a JPEG-compressed thumbnail — typically 160×120 — inside the same APP1 segment. If the photo was cropped, annotated or redacted before export, the thumbnail can still show the earlier framing. Removing the metadata removes the thumbnail with it, because both live in the same segment.
Canon, Nikon, Sony and Apple write private tag ranges with lens corrections, focus distance, shutter count and sometimes serial numbers. This tool does not decode them, so they never appear as rows in the report, but they sit inside the same APP1 segment and are removed with it. A report that lists nine fields can still be removing eleven.
The compressed image data lives in separate segments (APP1 and the rest are ancillary). That is why this tool can drop every metadata segment and copy the scan data byte for byte: no re-encoding, no generation loss, no colour shift. The two deliberate exceptions are the ICC profile and the orientation tag.
Drop a JPEG, PNG or WebP file into the panel. Parsing happens in a Web Worker; nothing is uploaded.
Compare the reported tags with the reference above. Unknown tags are removed even when they are not listed — the report shows what identifies you, not everything in the file.
Download the cleaned copy and open it in your usual viewer to confirm it still looks right. The original is untouched.
Before you rely on it
Both tags change how the pixels are interpreted. Strip the orientation tag and a portrait photo displays sideways in viewers that trust the file; strip the ICC profile and colours shift, sometimes badly on wide-gamut displays. They reveal nothing about where you were or who you are, so they are kept deliberately — and this is the one place where “remove everything” would make the file worse.
EXIF is the camera-written block described above: capture settings, device, GPS. XMP is an XML vocabulary that can hold descriptions, ratings, edit history and AI labels. IPTC is an older news-industry set of fields, now usually carried inside XMP. All three are removed by this tool; only EXIF has the tag structure on this page.
Yes, and it is common. Messaging apps, social platforms, screenshot tools and many export presets strip the block entirely. Front-facing cameras often skip GPS even when location is enabled.
No. The metadata segments are removed and the compressed image data is copied unchanged. The file gets slightly smaller because the metadata is gone; the pixels are identical.
Usually not — both re-encode and drop the block. But “usually” is doing work in that sentence, and it depends on whether the file was sent as a photo or as a document. Inspect the copy you actually hold rather than assuming.
It was inside the APP1 segment and was removed with it. That is intentional: a thumbnail is a second, often older, copy of the image, and it is invisible in a normal file browser.