Remove Gemini Watermark

The Gemini watermark is the small four-point sparkle icon Google adds to the bottom-right corner of every image its Gemini model generates — including images from the Gemini 2.5 Flash Image model the community nicknamed “Nano Banana.” It exists to signal that a picture is AI-generated. Unlike a hand-placed signature or a stock watermark that can land anywhere in the frame, this mark is rendered into the same fixed position on every single output.

That predictability is the whole reason one-click removal is possible. On a general remover you brush the exact region to clean and an AI labeller works out what sits underneath. Here there’s nothing to identify and nothing to point at: the tool already knows the sparkle is in the bottom-right corner, so it sends the AI a fixed instruction to rebuild it. You skip the brushing step entirely — upload, tap Remove, and the mark is gone.

The one-click flow above is tuned for the standard case: a sparkle sitting in the bottom-right corner, which covers the vast majority of Gemini and Nano Banana exports. But there are a few exceptions. Some older Gemini versions placed the mark bottom-left or slightly inset from the edge, and if you’re working from a screenshot it can be cropped to almost any position — or sit next to a logo or app UI.

For those cases, the general Watermark Remover is the right tool. There you brush over the exact spot, so it can clear the mark in any position, plus the Gemini logo in a screenshot or anything else on the image. Same AI, same free daily quota — the only difference is that you paint the region instead of letting the fixed bottom-right instruction handle it. Between the one-click flow here and the brush flow there, every Gemini watermark is covered.

It’s worth being precise about what this tool does and does not touch. Google applies two separate things to a Gemini image. The first is the visible Gemini watermark — the sparkle in the corner — which is what gets rebuilt. The second is SynthID: an invisible, pixel-level signature spread across the entire image and designed to survive crops, re-encodes, and colour edits.

Clearing the sparkle cleans the visible corner and nothing else, so the SynthID layer in the rest of the image remains intact. That’s by design. The point of the tool is visible cleanup for your own workflow — a tidy image for a presentation, a thumbnail, or a personal project — not hiding a picture’s AI origin. Any service that claims to strip SynthID along with the visible mark is over-promising; the two are different problems.

For personal or non-deceptive use, removing the Gemini watermark is generally fine. Google’s terms give you broad rights over content you generate with Gemini, so cleaning up your own images for your own purposes is well within bounds. The grey area is about context, not the act of removal itself.

Where it gets sensitive is anywhere viewers would reasonably expect to know an image is AI-generated: news and journalism, stock marketplaces that require AI disclosure, design competitions that ban AI entries, or any setting where a hidden origin could mislead. Removing the mark in those contexts can break a platform’s rules even if it breaks no law. The honest rule of thumb: clear it freely for your own workflow, and disclose the image’s AI origin honestly when it leaves your hands.