Google image checks
Gemini Image Detector for Nano Banana AI Photos
Use this Gemini image detector when a picture may have come from Google image generation, including the tools known as Nano Banana. Slop or Not reads Google SynthID watermarks, signed credentials, and editable metadata, then scores the pixels with a classifier trained on millions of human and AI images when no readable provenance or watermark resolves the file. Online checks stay away from outside AI companies.
The only AI image detector that reads Google DeepMind's SynthID watermark from both Google and OpenAI images.
Google's own detector checks only Google's images. OpenAI's checks only OpenAI's. Slop or Not reads both.
What it checks
Check Nano Banana and Gemini Images for SynthID Signals
Google SynthID watermark signals plus any C2PA/IPTC metadata
Generator traces in image texture, edges, and repeated details
Fallback classifier output for copied, exported, or recompressed files
Good Uses
- Checking a Gemini image before using it in class materials
- Reviewing suspicious profile photos from search or messaging apps
- Testing whether a file still carries readable provenance
Limits to Know
- A missing SynthID read is not proof of a real photo, so the image model still scores the pixels
- SynthID is built to survive ordinary sharing, but heavy edits can still make it inconclusive
- Mixed human edits can make the final verdict less direct
FAQ
Gemini AI Image Detector Questions
What is Nano Banana?
Nano Banana is the community name for Google's Gemini image generation model. Google says images generated with Gemini include an invisible SynthID watermark, and Slop or Not checks that signal before judging the pixels.
Can Slop or Not read Gemini provenance?
Yes. Slop or Not checks SynthID, Google's invisible watermark for generated media, and reads C2PA/IPTC metadata when the file exposes it. If those signals are missing, it uses visual image detection.
Can Gemini images lose their watermark?
Screenshots can remove metadata, but SynthID is embedded in the image and is not easy to strip. Repeated heavy edits, severe recompression, or very small copies can still make the check inconclusive.
Is this a different checker from the AI image detector?
No. Every Slop or Not image check works the same way: SynthID watermarks, C2PA and IPTC credentials, and the image model when neither settles the file. Gemini and Nano Banana files are the case where Google SynthID is usually the evidence that decides it, so the signals, limits, and answers here are written for those files. Upload the same picture anywhere on the site and the verdict is the same.
What counts as a supported provider for SynthID?
Slop or Not reads image SynthID from Google/Gemini and OpenAI images today. Other generators, plus files where SynthID was never applied or was stripped, fall back to C2PA/IPTC metadata and then the pixel classifier.
Run the Next Check on Your iPhone or Mac
The Slop or Not app checks images and text on-device, so nothing you check ever leaves your device.
Related checks
Compare Related Detector Workflows
SynthID Detector
Free SynthID detector: check an image for Google and OpenAI SynthID watermark signals online, or run the check on-device on iPhone and Mac.
AI Image Detector
Use Slop or Not as an AI image detector when the source is unknown. It checks SynthID, C2PA/IPTC metadata, and pixels before reporting a result.
AI Watermark Detector
Use this AI watermark detector and checker to read image SynthID, verified C2PA Content Credentials, and IPTC metadata before any pixel check.