Grok image detector
Fewer than 1 in 100 original camera photos are wrongly flagged (0.6%)
Can Perspicax detect Grok images?
Yes. Perspicax caught 96.2% of Grok Imagine 2.0 images in testing. Grok Imagine is xAI's image model, used in the Grok apps and on X. xAI has published no official statement on labels or watermarks for its images, so don't count on finding one; Perspicax checks the pixels.
Where these images come from
People make Grok images at grok.com/imagine and in the Grok and X apps. Developers use xAI's API, where Grok Imagine Image 2.0 is the recommended model.
Labels and watermarks
xAI's image documentation says nothing about provenance. Claims that Grok images carry Content Credentials, a visible wordmark or prompt metadata are unconfirmed, and xAI is not a confirmed SynthID user. X also strips file metadata on upload.
- Content Credentials (C2PA)Live
- SynthID watermark checkSupported
What we measured
- Grok Imagine 2.096.2%
Worth knowing
Grok images are often made and shared directly on X. To check one, save the image or copy the image's address, then drop or paste it into Perspicax.
How it worksQuestions about Grok images
Does Grok label its images?
xAI has published no official provenance statement for Grok images, so there may be no label to find. Perspicax checks the pixels, so it doesn't need one.
Can I check a Grok image posted on X?
Yes. Save the image or copy the image's address (not the post link), then drop or paste it into Perspicax. X strips file labels, which is why the pixel check matters.
Does SynthID detect Grok images?
Not that we know of: xAI isn't a confirmed SynthID user. Perspicax's pixel check doesn't depend on a watermark.
Which Grok model did you test?
Grok Imagine 2.0: 96.2% of its images caught, on copies collected after release and mostly re-saved as 1 MP WebP.
Before you believe it, check it.
Your first scan is free, no account needed. A free account gives you 2 scans every day.