Watermark

A watermark is a visible or embedded mark added to an image or video to indicate its ownership, source, or that it was generated by AI.

Watermarks come in two forms. A visible watermark is a logo, text or pattern overlaid on the content. An invisible watermark is data woven into the pixels themselves — undetectable to the eye but readable by software designed to verify provenance.

How it works

A visible watermark is composited onto the output, often semi-transparent and tiled or placed in a corner. It is meant to be seen, so the source is clear even when the file is shared.

An invisible or forensic watermark works differently. It makes tiny, statistically structured changes to pixel values that survive resizing, compression and screenshots but stay imperceptible. Detection tools then read the hidden signal to confirm where the content came from. Provenance standards such as C2PA attach signed metadata about how a file was made, and systems like Google's SynthID embed detectable markers specifically to flag AI-generated media. Together these approaches support disclosure and authenticity efforts as AI content becomes common.

Why it matters

For creators, a visible watermark on trial output is a barrier — you cannot use the asset cleanly for business until you upgrade. For the wider ecosystem, invisible watermarking and provenance metadata are becoming central to transparency: platforms, regulators and publishers increasingly want a reliable way to tell whether media is synthetic. Understanding the difference helps you choose tools that deliver clean, usable files while respecting disclosure norms.

In eaxy

eaxy is a paid studio with no free tier, so your exports are clean and ready to use — there are no promotional watermarks stamped across your images or video. Commercial use is covered on Pro and higher plans, meaning the files you download go straight into your work rather than serving as watermarked previews.

Related terms

• Commercial License
• Render
• Generative AI