How to Remove a Leonardo AI Watermark and the Metadata It Ships With

By Calabi Labs Editorial Team · 2026-08-02

Leonardo.ai images carry more than a corner watermark — the file itself is stamped with AI provenance metadata that survives cropping. Crop the visible mark yourself, then run the image through Calabi to strip the invisible C2PA and XMP AI flags and rewrite the file so it reads as a normal phone photo. No tool can guarantee a platform won't flag you; results vary by platform and source model.

What actually flags your file

The watermark you can see is the smallest part of the problem. A Leonardo export ships with a metadata layer platforms actually scan: C2PA / Content Credentials manifests stored as JUMBF atoms, XMP flags such as DigitalSourceType: trainedAlgorithmicMedia, and generator or tool tags naming the software that made the image.

That layer is invisible and it survives editing. Crop the corner, resize the image, screenshot it — the manifest and the AI flags can ride along in the new file. Detection systems read those tags, not the pixels.

A raw AI export can carry well over a hundred metadata tags. A real phone photo carries a lean, structural set. The gap between those two profiles is one of the easiest things for a platform to notice.

How Calabi handles it

Calabi runs one automatic pass in three stages. It STRIPS the signals platforms scan for: the C2PA / Content Credentials manifest, the XMP AI flags like trainedAlgorithmicMedia, and the generator tags. It INJECTS an authentic phone-capture identity — Make, Model, Software version, GPS coordinates, and a capture timestamp — using device profiles like iPhone 17 Pro Max, iPhone 16 Pro, Pixel 8 Pro, or Galaxy S24 Ultra. Then it VERIFIES with a forensic proof card from the same ExifTool scan platforms use, so you see exactly what changed.

In testing, Calabi has reduced 18 JUMBF/C2PA atoms to 0, cleared 16 C2PA references, removed the trainedAlgorithmicMedia flag, and trimmed a raw AI export's 144 metadata tags down to roughly 94 neutral structural ones — all confirmed with ExifTool.

For the visible Leonardo watermark itself: crop it out, then run the cropped file through Calabi. Calabi does not erase logos pixel by pixel — it removes the invisible layer the crop can't reach.

Steps:

Crop the visible Leonardo watermark out of the image if it has one.
Upload the image to calabilabs.com.
The pipeline automatically strips the AI metadata and injects a phone-capture identity.
Read the forensic proof card to confirm the C2PA atoms and AI flags are gone.
Download the cleaned file.

FAQ

Does Calabi erase the Leonardo watermark from the picture?

No. Calabi does not paint over or reconstruct any part of the image. Crop the visible mark yourself, then use Calabi to strip the invisible C2PA and metadata layer that a crop leaves behind — that layer is what usually triggers a flag.

Does processing change how the image looks?

No. Calabi works on invisible signals and metadata only. It does not recolor, edit, or alter any region of the picture — what you see stays the same.

How do I know it actually removed the AI flags?

Every run ends with a forensic proof card generated by ExifTool — the same scanner platforms use. It shows the C2PA atoms and the trainedAlgorithmicMedia flag reduced to zero before you download.

Try Calabi free at calabilabs.com — 10 cleans, no card.