Extract Image Watermark
Free invisible image watermark reading/detection API. Extracts the marker previously embedded in an image and tells you whether an image carries a watermark at all. Accepts file upload, URL, or Base64. Useful for content provenance, theft forensics, and AIGC label verification. · content provenance verification, theft forensics, AIGC label verification, originality confirmation, distribution-channel tracking · Extract Image Watermark
free invisible watermark reader API, blind watermark extraction API, watermark detection API, image provenance verification, digital watermark extraction, extract watermark api
POST /watermark/decode - Extract Image Watermark
Found an image you suspect was stolen or AI-generated, and want to check whether it carries a hidden marker? Hand it straight to this endpoint and it pulls the embedded marker back out in full.
Overview
This endpoint detects whether an image contains an invisible watermark and, if so, recovers the original marker that was written in. It still reads back reliably in most cases even after the image has been compressed, cropped, or forwarded many times. It also guards against false positives: if the image really was never processed, it clearly returns no result instead of stitching together something meaningless.
Notes
If you changed any advanced parameters when embedding (such as ecc or modeltype), you must pass the same values here when reading. If embedding used the defaults, just leave these empty.