Turning the visible watermark off
In the Gemini app or on the web, open your profile picture or the menu, go to Settings, choose Media watermark, and switch the visible watermark off. Images you create from then on come without the sparkle mark in the corner (TechRadar, TechCrunch).
The setting doesn’t reach back: images you already saved keep their mark. To get one without it, make the image again with the setting off.
Where it stays on anyway
In India, South Korea and Vietnam the visible watermark is applied automatically and the setting doesn’t turn it off, because rules there require AI-generated images to be labeled. India’s 2026 IT rules, for example, call for prominent, persistent labels on synthetic media. The EU’s AI Act also requires AI-generated content to be marked in a machine-readable way from August 2026.
What stays in the image
SynthID. Google embeds an invisible watermark in the pixels of every image its models generate. It is designed to survive cropping, resizing and compression, and turning off the visible mark doesn’t remove it — so Google’s own tools can still tell the image was made with its AI.
Metadata. AI images can also carry a label in their metadata — an IPTC “digital source type” saying the image was generated, or signed Content Credentials (C2PA) recording how it was made. The photo metadata viewer shows both when they are there. It can’t see SynthID, which lives in the pixels rather than the metadata.
Why this site doesn’t remove watermarks
FixMyFiles doesn’t offer a watermark remover, for AI images or anyone else’s. A watermark on someone else’s photo is usually there to protect their copyright, and an AI label is increasingly required by law where the image is published. What the site does offer runs the other way: adding your own watermark to your photos, and seeing what a photo says about itself before you share it.