Google's SynthID Detector (official)
Google adds an invisible watermark called SynthID to images made with its AI, including Gemini and Nano Banana. You can't see it, and it is built to survive common edits such as resizing and compression.
On 7 October 2026, Google opened its SynthID Detector to everyone. Before that, it was limited to a waitlist of journalists and researchers.
- Go to synthid.com.
- Sign in. An account is required.
- Upload the image. It also takes video and audio.
- Read the result. It tells you whether it found a SynthID watermark.
Worth knowing before you rely on it:
- It only finds SynthID. It is not a general AI detector. Reports say it also covers some partners that use SynthID, but not every AI tool.
- It cannot tell an image that was fully made by AI from a real photo that was edited with AI.
- It works in English only for now, as of October 2026.
- You upload the image to Google to run the check.
Ask the Gemini app
Google has also built the same check into Gemini. Upload the image in a chat and ask something like "Was this image made with Google AI?" Gemini looks for SynthID and answers. Google has said similar checks are available through About this image in Search and in Google Lens.
This is handy on a phone, where you may already have the Gemini app.
Check the file's Content Credentials
Gemini also attaches Content Credentials (also called C2PA) to its images. This is information in the file that says it was made with Google AI. Unlike SynthID, it is plain data that anyone can read, and that is what our tools look at.
- Open What Made This or the metadata viewer.
- Add the image. It is read on your device and not uploaded.
- Look at the result. If the file has Content Credentials from Google, you will see it named as the source.
Content Credentials are often lost along the way. Many social apps and chat apps drop them when you upload or forward an image, and people can remove them on purpose (we offer a C2PA remover for your own files). So this check is strongest on an original download and weakest on an image you found on social media.
Look for the visible sparkle
Images from the Gemini app have long carried a small sparkle near the bottom-right corner. If you see it, the image very likely came from Gemini. But its absence means little:
- Since August 2026, many users can switch the sparkle off for new images.
- Images made through Google's developer tools may not have it.
- It can be cropped out or removed, for example with our Gemini watermark remover.
If you just want to know what the sparkle on Nano Banana images is, see the Nano Banana 2.1 watermark guide.
What "no result" really means
Every one of these checks can come back empty, and an empty result is not proof that a person made the image. It can mean:
- The image came from an AI tool that uses neither SynthID nor Content Credentials.
- The file data was stripped by an app or on purpose.
- The image was screenshotted, heavily edited or re-encoded.
- It really is a normal photo.
A positive result is much more meaningful than a negative one. If several checks agree, you can be fairly confident. If none find anything, the honest answer is "unknown".
What UnmarkPro cannot do here
- It does not detect or read SynthID. Use Google's SynthID Detector or the Gemini app for that.
- It cannot tell you an image is "real". It only reports what the file's data says.
- If the data has been removed, it will say the source can't be determined rather than guess.
Frequently Asked Questions
What is the most reliable way to check if an image was made by Gemini?
Google's own SynthID Detector at synthid.com. It looks for the invisible SynthID watermark that Google adds to images made with its AI. Since 7 October 2026 anyone can use it for free after signing in.
Can UnmarkPro tell me if an image has SynthID?
No. UnmarkPro does not read or detect SynthID. Our metadata viewer and What Made This tool read the data attached to the file, such as Content Credentials, and show what it says. For SynthID, use Google's tools.
Can I ask the Gemini app if an image was made by Google AI?
Yes. Google says you can upload an image to Gemini and ask whether it was made with Google AI. Gemini checks for SynthID and answers in the chat.
If a checker finds nothing, does that mean the image is real?
No. A no-result only means that this particular check did not find its signal. The image may come from a tool that does not use SynthID or Content Credentials, the file data may have been removed, or the image may have been changed a lot.
Does the visible sparkle prove an image came from Gemini?
It is a strong hint, but not proof. The sparkle can be added to other images, and since August 2026 many Gemini images are made without it. Use it together with the other checks on this page.
Sources: reporting by MacRumors and PetaPixel on the SynthID Detector (October 2026), Search Engine Journal (August 2026), and Google's Gemini Apps Help page. Features and availability can change.
Related Guides & Topics
What Is SynthID? Google's Invisible Watermark
What SynthID is, where it is used, and why it cannot be removed.
Instagram "AI Info" Label: Why It Appears
Check a photo before posting and what to do about a label.
How to Turn Off the Gemini Watermark in Settings
The Media watermark setting, who can use it, and what it leaves behind.