How to Tell if an Image Was Made With Google Gemini
Google marks its images more thoroughly than most. Every Gemini and Nano Banana output carries SynthID, an invisible watermark in the pixels, and newer models add Content Credentials on top. Check the credentials first from the original file; if they are gone, the SynthID watermark is designed to survive screenshots and re-uploads. When neither is available, fall back to a detector and the visual checks.
What Google adds to an image
SynthID, from Google DeepMind, is a watermark woven into the pixels. Google applies it to all Imagen and Gemini image output, and a public detector has been available in the Gemini app since late 2025. Because it lives in the image, it is built to hold up through screenshots, resizing, and compression, unlike metadata.
Content Credentials are signed C2PA metadata. Since the Gemini 3 Pro image models in November 2025, new output carries both SynthID and Content Credentials. In May 2026 Google announced that Chrome and Search will surface C2PA checks, so the credentials are becoming visible to ordinary users as well as tools.
How to check a Gemini image
Start with the original file. Inspect it for Content Credentials at contentcredentials.org/verify, where a valid record names how the image was made. For the watermark, upload the image to the Gemini app, which can report whether SynthID is present. The catch is access: SynthID is readable only through Google’s detector, so it helps platforms and Google’s own tools more than an open inspector.
What survives a screenshot
A screenshot drops the Content Credentials but keeps the SynthID watermark. That is the opposite of the OpenAI case only in emphasis, since both vendors now use the same two-layer approach. The practical rule is the same: metadata is easy to lose, a pixel watermark is harder. If you can only get a screenshot, the watermark is the signal left to check, and only Google can read it.
Detector and visual checks as backup
If both marks are missing, score the original file with the free AI image checker and treat the result as a probability, not proof. Then run the visual checks for hands, text, reflections, and background logic. Nano Banana is strong at instruction following and text rendering, so the old text and hands tells are weaker than they were. Our overview of detector limits is in Do AI Image Detectors Actually Work?, and if the image came from OpenAI tools instead, see the DALL-E and ChatGPT guide.
Common questions
What is Nano Banana?
Nano Banana is the nickname for Google's Gemini image generation and editing model. Output is produced through Gemini, and every image carries Google's SynthID watermark.
Do Google Gemini images have C2PA Content Credentials?
Newer models do. Since the Gemini 3 Pro image models in November 2025, new output carries both SynthID and Content Credentials. Google also announced in May 2026 that Chrome and Search will surface C2PA checks for consumers.
How do I check an image for SynthID?
SynthID can only be read by Google's detector, which is available in the Gemini app for images you upload there. Unlike C2PA, there is no open inspector, so a normal viewer cannot read it directly.
Does SynthID survive a screenshot?
It is designed to. SynthID is embedded in the pixels and built to hold up through screenshots, resizing, and compression, which is the opposite of the C2PA metadata layer. That makes it the more durable of the two signals.
Check any image in under 3 seconds
AI Detector scores an image and never stores it.
