Besides the Content Credential, every image and video you create with Samsa carries a second, independent marking layer: an invisible watermark. It's a subtle, machine-readable pattern woven directly into the pixels — you can't see it, and it doesn't change how your content looks, but verification tools can detect it.
The Content Credential lives in the file's metadata. That makes it precise and rich in information — but metadata can get lost: many social networks and messaging apps strip it when you upload a file. The invisible watermark solves this. Because it's part of the pixels themselves, it stays in the content even when the metadata is gone. Together, the two layers make sure AI content remains identifiable across the everyday life of a marketing asset.
This two-layer approach — signed metadata plus an imperceptible watermark — is also what the European Commission's Code of Practice on Transparency of AI-Generated Content describes for images and videos.
The watermark carries one simple message: this content was created with Samsa. It doesn't encode your prompts, your name, your organization, or any other personal data.
The watermark is designed to survive everyday handling of marketing assets:
Cropping and resizing
Compression (for example when a platform re-encodes your upload)
Screenshots
Format conversions (for example PNG to JPEG or WebP)
Mirroring and similar routine adjustments
One honest note: no watermarking technology in existence can guarantee survival against every deliberate, targeted removal attempt. That's precisely why Samsa uses multiple independent layers — the watermark complements the Content Credential rather than replacing it.
Nothing. The watermark is embedded automatically for every image and video, in the same step in which the file is signed. There is no setting to manage and no effect on visual quality.
