Skip to content
AI Origin CheckRun a free check

No fake AI score

AI Watermark Checker: Evidence, Metadata & Limits

Start with inspectable evidence: signed Content Credentials, ordinary image metadata, and visible text artifacts. Provider-specific signals require provider-supported detection.

Drop an image here

JPEG, PNG or WebP • up to 25 MB

Your file stays on your device. Analysis runs locally in your browser.

There is no universal AI watermark

AI systems do not share one public watermark. Some products attach C2PA provenance, some embed provider-controlled signals such as SynthID, some leave ordinary metadata, and some add nothing durable.

This tool validates available C2PA records and reports editable metadata as hints. It does not infer an AI percentage from pixels or prose.

Watermarked can mean processed—not wholly generated

A signal may indicate that a provider generated, edited, translated, upscaled, or otherwise processed part of an asset. Provenance actions and digital source types provide the context needed to avoid overclaiming.

Primary sources and further reading

For technical and policy detail, consult the C2PA specification, Content Authenticity Initiative developer documentation, Anthropic transparency material, Google DeepMind’s SynthID overview, and the European Commission’s AI transparency guidance.