Image metadata explained

“Metadata” covers at least three separate systems that happen to live in the same file. They were designed by different bodies for different purposes, and knowing which is which explains why some information travels and some vanishes.

There is also the format's own housekeeping — colour profile, density tag, and similar. That is not really metadata about the picture, but it lives alongside it and is subject to the same losses.

This is the practical question, and the answer disappoints people regularly.

• Re-encoding through a browser strips essentially everything. Any tool that decodes an image to a canvas and writes a new file — which includes every image tool on this site — produces a clean file with no EXIF, IPTC or XMP. It is a consequence of how the canvas API works, not a choice.
• Most social platforms strip metadata on upload. Partly for privacy, partly for size.
• Screenshots have none of the original's metadata, only their own.
• Copying a file preserves it. Moving, renaming and ZIPping do not touch the contents.
• Dedicated photo software preserves it deliberately, and lets you choose what to include on export.

EXIF from a phone routinely includes the exact coordinates where a picture was taken, along with the device and the timestamp. Publishing an unprocessed camera file publishes all of that.

• Check before posting anything taken at home, at a school, or anywhere you would not put on a map.
• Turn off location tagging in the camera app if you never want it.
• A re-encode strips it as a side effect — running an image through the format converter or the JPG compressor leaves a file with no GPS in it. Verify rather than assume if the stakes are real.

Agencies differ, and it is worth reading the one you submit to. Broadly:

• Embedded IPTC is often read on ingest. Title, description and keywords written into the file can pre-fill the submission form, which is the fastest route for a large batch.
• A CSV is the other route. Filename plus columns for title, description and keywords, uploaded alongside the images — which is what the bulk metadata generator exports.
• EXIF is rarely required but is sometimes used to verify a claim about how an image was made.

What goes in those fields is a separate craft — titles and keywords that hold up covers it, and the ranker reviews a set before you commit to it.

Since the browser canvas discards metadata, writing any of it back has to be done deliberately. The DPI converter does exactly that for one field: it inserts a real density value into the JPG's JFIF header or the PNG's pHYs chunk after re-encoding, so the print size a layout application reads is the one you asked for.

WebP has no density field at all, which is why the DPI tool offers only JPG and PNG as outputs. What that value does — and does not — mean is in what 300 DPI means.

• Does this file contain a location I would not publish?
• Does it carry my copyright and credit, if it is going somewhere that respects them?
• Is the file I am uploading the delivery copy, with the master kept intact elsewhere?
• If it is going to print, does it carry the density the printer asked for?

A note on what these tools can see

The image tools on this site never upload anything — every file is decoded and re-encoded in your own browser, so no metadata reaches any server, including ours.