Skip to content
Delete Image Metadata

Grok Image Metadata

Grok marks its images with a visible watermark, and xAI's rules forbid stripping provenance from them. See what your file carries — the checker reads it and changes nothing.

  • .jpg
  • .jpeg
  • .png
  • .webp
Visible
Grok's watermark, with no setting to turn it off
26 Jun 2026
xAI policy barring removal of provenance and watermarks
Read-only
This checker changes nothing in the file
Processed in your browser

How it works

  1. 1

    Load the Grok image

    Read in your browser; nothing is uploaded.

  2. 2

    See what it carries

    Any C2PA credential, AI declaration or generator name is shown.

  3. 3

    Disclose where it applies

    On X, use the Made with AI option rather than relying on the automatic label.

How Grok marks its images

Images and videos from Grok carry a visible Grok watermark, and xAI's own FAQ says there is no setting to remove it.

Whether Grok images also carry a signed C2PA manifest is disputed: some reports describe credentials in every export, others find none. The file you have settles it — load it above.

xAI's rules

xAI's Acceptable Use Policy, effective 26 June 2026, prohibits "stripping, altering or circumventing embedded provenance metadata or watermarks". That covers both the visible mark and any metadata a Grok image carries.

So this page reads Grok files and does not change them, and this site does not remove visible watermarks from any image.

Grok images on X

X introduced a "Made with AI" label in early 2026, which you can add yourself and which X also applies automatically to some AI images. A label X applies automatically cannot be removed by the poster.

X has not published what triggers the automatic label, and one published test found a Grok image left unlabelled while a ChatGPT image was labelled. An unlabelled post is not a pass: X has announced penalties for creators who post undisclosed AI images of armed conflict, so use the disclosure option when it applies.

What this tool does

  • Settles whether your Grok image carries a credential
  • Explains how X labels Grok images, and where it doesn't
  • Read-only: nothing in the file is changed

What it does not do

  • Does not remove metadata or watermarks — xAI's policy forbids it
  • Cannot remove or detect marks in the pixels
  • Cannot predict X's automatic label, whose triggers are unpublished

Grok Image Metadata

Grok marks its images with a visible watermark, and xAI's rules forbid stripping provenance from them. See what your file carries — the checker reads it and changes nothing.

Open the tool

Reviewed and updated . Behaviour described on this page is covered by the project’s automated tests.

Frequently asked questions

Do Grok images have a watermark?

Yes. Grok adds a visible watermark to generated images and videos, and xAI says there is no setting to remove it.

Does Grok add C2PA metadata?

Reports disagree. Load your Grok image above and the checker shows whether that file carries a signed credential.

Can I remove the Grok watermark or metadata?

xAI's Acceptable Use Policy prohibits stripping, altering or circumventing provenance metadata or watermarks. This site does not remove either from Grok images.

Why didn't X label my Grok image as AI?

X has not published what triggers its automatic label, and a published test found Grok images unlabelled. That leaves disclosure to you.

Do I need to label Grok images on X?

Use X's Made with AI option when posting AI images, especially realistic ones. X has announced penalties for undisclosed AI images of armed conflict.

Keep going

Related tools

Back to the main metadata remover