Skip to content

Gemini and Imagen image detector

Google attaches provenance information to output from Imagen, Gemini and its consumer image tools, including C2PA Content Credentials on several surfaces. Where that data survives, identifying the source is straightforward.

One important limitation: Google also applies SynthID, an invisible watermark encoded into the pixels themselves. SynthID survives cropping and re-encoding, but it can only be read with Google's own detector — no third-party tool, including this one, can check it. What this tool reads is the metadata layer.

Drag a file here, or

JPEG, PNG, WebP, AVIF, PDF or .txt · up to 10 MB · you can also paste a screenshot

By uploading you agree to our Terms and Privacy Policy. Your file, IP address, approximate location and device details are stored.

What this checks

  • C2PA manifests naming Google, Imagen or Gemini
  • EXIF and XMP fields identifying Google's image tools
  • IPTC trainedAlgorithmicMedia assertions
  • Image classifiers where the metadata has been stripped

Frequently asked questions

Can you detect SynthID?

No. SynthID is a pixel-level watermark readable only by Google's own detector. Any third-party tool claiming to read it is not doing what it says.

Does Gemini always mark its images?

Google applies SynthID broadly, and Content Credentials on several surfaces. But the metadata layer is removed by the usual culprits — screenshots, re-encoding, platform uploads.

What if the image was downloaded from social media?

Expect the metadata to be gone and the result to rest on the classifiers, with a wide uncertainty range to match.

Related checks