Watermark Detector
Scan an image for visible watermarks, semi-transparent overlays and stock-photo metadata, and get a confidence score.
Drag & drop an image here, or click to browse
No uploads. Your files stay on your device.
Signals checked
How it works
Drop an image and it is decoded straight into a canvas in your browser, then screened for the fingerprints a visible watermark leaves behind: detail concentrated in one corner or across the centre of an otherwise smooth photo, pixels that are all lightened or darkened by the same uniform amount (the signature of a semi-transparent overlay), a pattern that repeats on a regular grid like a tiled stock-agency stamp, and partially transparent pixels in a PNG or WebP overlay. In parallel the raw file bytes are searched for stock-agency, licence and provenance strings such as Shutterstock, Getty, C2PA content credentials or AI-generation markers.
Each signal is scored, the scores add up to a confidence value out of 100, and the sensitivity selector lets you re-judge the same scan instantly without re-reading the file. This is a heuristic screening tool, not forensic proof: a busy photo can look suspicious, and truly invisible watermarks — Digimarc, Google SynthID and other steganographic marks — cannot be decoded in a browser, so a clean result never guarantees an image is unmarked. Nothing is ever uploaded; the scan, the report and the download all happen on your device.
Frequently asked questions
How do I know if an image has a watermark?
Drop the image on this tool and it scans the pixels for the four things a visible watermark leaves behind: detail concentrated in one corner or the centre of an otherwise smooth photo, pixels shifted lighter or darker by one uniform amount (the signature of a semi-transparent overlay), a pattern that repeats on a regular grid like a tiled stock stamp, and partly transparent pixels. At the same time the raw file bytes are searched for stock-agency and licence strings such as Shutterstock, Getty or C2PA content credentials. Each match adds to a confidence score out of 100 and the suspicious areas are outlined on the preview.
Can it detect invisible or hidden watermarks?
No, and no browser tool honestly can. Invisible watermarks such as Digimarc, Google SynthID and other steganographic marks are embedded in the frequency domain with a secret key and can only be read by the vendor's own detector. This tool can tell you when a file's metadata mentions Digimarc, SynthID or C2PA content credentials, which is often the practical giveaway, but a clean result only means no visible watermark and no telltale metadata were found — it is not proof the image is unmarked.
Is my image uploaded anywhere?
Never. The file is decoded into a canvas and the metadata scan reads its bytes with the browser's own File API, so every calculation happens on your device — you can disconnect from the internet after the page loads and the scan still works. That matters for client photos, screenshots and licensed material, which competing watermark checkers require you to upload to their servers. The report you copy or download is generated locally too.