How to view and remove image metadata

Load one image and the workbench lists the metadata it can read — camera and dates, GPS coordinates, IPTC records in a JPEG, XMP, PNG text chunks — and exports a re-encoded copy of the picture with none of it left. The file stays in this browser.

Measured here: a 7.1 KB JPEG listing 19 fields came back as a 5.9 KB file with no EXIF or XMP block at all; a 417 B PNG that carried no metadata still listed 7 structural fields and came back as 1.1 KB.

  1. Load one image. Choosing another replaces it, and a text file is refused with a message while the loaded picture stays.
  2. Press Process files. The table lists one row per field and the metrics report how many fields were found, the input size and the size of the file you can download.
  3. Read the table. Nested values keep their path: an XMP title arrives as title.lang and title.value, and a keyword list as subject.0, subject.1.
  4. Press Download result. The copy is the same picture re-encoded without metadata and is named after the source: photo-exif-gps.jpg became photo-exif-gps-metadata-removed.jpg.
  5. WebP, GIF and SVG are re-encoded too, but their metadata cannot be read: the run says so instead of listing fields. GIF and SVG come out as JPEG, WebP stays WebP.

What the table reads, what the download contains and what is refused

What the table reads

JPEG: EXIF (Make, Model, Orientation, XResolution, YResolution, Software, ModifyDate), GPS with its references and altitude plus the derived latitude and longitude, and IPTC-IIM records from the APP13 segment — a test file listed ObjectName, Byline, DateCreated and Caption.

PNG: the structural chunks (ImageWidth, ImageHeight, BitDepth, ColorType, Compression, Filter, Interlace) and tEXt chunks such as Software, Comment and Author. XMP in a JPEG listed title, creator and subject. A JPEG that carries no metadata is reported as such: the table stays empty, no field count appears and the picture is re-encoded anyway (a 425 B sample came back at 771 B).

What the download contains

The download is a new file, never the original. JPEG, PNG and WebP keep their format — JPEG and WebP re-encoded at quality 0.92, PNG lossless — and anything else the browser can draw is written as JPEG. Re-encoding moves the size: a 12,000×9,000 JPEG of 32.7 MB came back at 44.7 MB, while the 7.1 KB EXIF sample came back at 5.9 KB.

No EXIF or XMP block survived in any output we measured. Colour: the JPEG encoder adds a small sRGB profile (456 bytes) the input did not have, so check wide-gamut pictures before publishing. Transparent areas stay transparent, and PNG output keeps an alpha channel.

Orientation, damaged files and refusals

An EXIF Orientation tag is applied to the pixels: a 100×60 file tagged Rotate 90 CW listed that one field and came back as a 60×100 picture with the tag gone.

A file the browser cannot decode — an empty file, for instance — is refused with a message, and pressing Download before a run answers with the reminder to process an image first. Removal is all-or-nothing: the whole picture is written again, so single fields cannot be kept.

Recent tools: