Why Imagine adds a Grok watermark

Generated images and videos include a Grok watermark so viewers can tell the media came from AI. In some places, labeling AI-generated content is also required by law.

There is no settings toggle to turn the watermark off.

What the policy bans

xAI's Acceptable Use Policy forbids stripping, altering, or circumventing embedded provenance metadata or watermarks. That covers both the visible mark and other provenance signals on outputs.

Support will not walk you through removal. Tools or prompts aimed at scrubbing the mark sit outside allowed use and can put the account at risk.

What you can do instead

Plan the composition so the mark sits in a quiet corner, or crop for a layout where the mark stays visible on the full export and a marked derivative is fine for the channel you ship to. Keep the original marked file when a client or platform asks for provenance.

If a platform needs unlabeled media, pick a workflow that platform allows under its own rules. Grok Imagine will keep labeling its outputs.