CheckSynthID

CheckSynthID is an independent hub for practical content provenance, focused on two different signals: C2PA Content Credentials and Google DeepMind’s SynthID watermark.

For C2PA, use the free Content Credentials reader to check whether a JPEG, PNG, or WebP carries signed Content Credentials. The reader runs in your browser, requires no account, and processes files locally—so there’s no upload queue. After you drop or choose an image (or try a sample), you’ll see a clear result such as Valid, Invalid, or None, and you can view the signer when credentials are present. For best outcomes, start from the original or least-recompressed export.

CheckSynthID is also built for understanding how SynthID results should be interpreted in workflows. SynthID is an imperceptible watermark embedded in pixels/audio, and the C2PA reader does not detect it. The directory’s guides explain what “Detected,” “not detected,” and “unclear” mean for the official detector, and why negative watermark outcomes are not proof of human-made content.

This tool is for editors, newsroom intake, agencies, and brand teams who need fast, local provenance checks—especially when assets come from unknown or third-party sources. Since credentials can be missing, stripped, or incomplete, treat Content Credentials as evidence to investigate alongside other signals rather than a sole verdict. The C2PA reader is free; deeper SynthID checking uses the official detector flow.

What does the free CheckSynthID C2PA reader check for?

It checks whether a JPEG, PNG, or WebP contains signed C2PA Content Credentials. You’ll get a clear Valid / Invalid / None status, and the signer is shown when credentials are present.

Does using the C2PA checker upload my images to a server?

No. The reader processes supported image files in your browser and requires no account, so your files stay on your device.

If the result is “None,” does that prove the image is human-made or AI-made?

No. “None” means Content Credentials aren’t present (or can’t be validated), which is a prompt to investigate further rather than a definitive authorship conclusion.

Can I use CheckSynthID’s C2PA reader to detect SynthID watermarks?

No. The C2PA reader validates signed metadata, while SynthID detection requires the separate SynthID detector workflow covered in the guides.

Is “not detected” from SynthID evidence that the content is human-made?

No. SynthID outcomes need careful interpretation, and “not detected” (or unclear results) doesn’t reliably prove human authorship by itself.

Is CheckSynthID affiliated with Google or the official SynthID Detector?

No. CheckSynthID is independent and does not run SynthID detection; it provides free C2PA tooling and SynthID education to help you read results correctly.