Remove Image Metadata

Create a pixel-only PNG copy of a browser-readable image.

One image up to 25 MiB and 24 million decoded pixels. Processing is local; the original file is unchanged.

No file ready? Download a sample, then select it below.

Make a new PNG from visible pixels

This operation decodes the source and writes a PNG from the rendered pixels. It does not copy the original file's EXIF, GPS, XMP or other attached blocks, and does not edit the source in place.

Create and check a copy

  1. Choose the original image.
  2. Press Create PNG copy and download the new PNG.
  3. Inspect the output with Image Metadata Viewer or another inspector.
  4. Review what the picture visibly reveals before sharing it.

Example: GPS in a phone JPEG

A phone photograph may contain location in EXIF. The new PNG is produced from decoded pixels rather than the original JPEG structure, so its original EXIF block is not copied. An address visible on a photographed sign remains visible.

Compare a tagged sample with its PNG copy

Download the synthetic EXIF JPEG. In Image Metadata Viewer, it reports Make: All Device Test, Model: Demo Camera and Artist: Synthetic sample. These are demonstration tags written onto generated shapes; no real camera or GPS information is used.

Select that JPEG here and press Create PNG copy. Expect a 640 × 480 PNG with the same visible shapes. Inspect the downloaded PNG with the metadata viewer: the original Make, Model and Artist tags should be absent. An empty or null metadata result alone is not a universal privacy certificate; this check verifies those known tags in this sample.

The filename is separate from embedded metadata. A download name can still reveal information when you share it, and this tool does not inspect visible faces, signs or text. Review the final picture as well as its parsed fields.

Is this complete privacy sanitization?

No. It does not blur faces, redact text, erase watermarks or detect hidden information in every possible form. The browser encoder may add its own format metadata. Removing original attached blocks is different from proving that a picture reveals nothing sensitive.

Why is the result larger or no longer animated?

PNG can use more bytes than a camera JPEG, and canvas records one rendered image rather than an animation. It also does not promise the original color-profile behavior in every viewer.

Inputs must fit 25 MiB and 24 million decoded pixels. Processing is local. For smaller output afterward, try Image Compressor, then inspect the final file again.