Watermark checks
SynthID Detector and Watermark Checker
Check any image with this SynthID detector. Slop or Not reads image SynthID from both Google and OpenAI images. The mark lives in the pixels, not metadata, so it survives screenshots and ordinary sharing, though heavy edits or severe recompression can still make a check inconclusive. When no mark or other readable provenance resolves the picture, a classifier trained on millions of human and AI images scores it instead. Online checks stay away from outside AI companies.
A SynthID checker finds stamped images. Slop or Not also checks the ones nobody stamped.
It reads SynthID from Google and OpenAI images, verifies C2PA Content Credentials, reads generator fingerprints, and scans the pixels.
- You can testPhotos, screenshots, product images, receipts, and profile pictures up to 8 MB.
- ResultA probability verdict, AI score, and source details when the file carries them.
- PrivacyChecked on hardware we own and run. We never pass your photo to OpenAI, Google, or any other AI company.
What it checks
Reads SynthID From Google and OpenAI Images, Then Checks the Rest
Google DeepMind's SynthID watermark in images from Google and Gemini
The same SynthID watermark in images from OpenAI, read with its own calibrated threshold
Fallback image-model score when no readable signal remains
Good Uses
- Checking a suspected OpenAI, Gemini, or Google-generated image
- Testing whether an exported image still carries watermark evidence
- Explaining why provenance and visual detection can disagree
Limits to Know
- SynthID exists only on images from generators that stamp it, so other generators are checked through credentials and the image model
- Heavy edits, severe recompression, or very small images can still make a check inconclusive
- A missing readable signal is not a clean bill of authenticity
FAQ
SynthID Detector Questions
Is SynthID the same as C2PA?
No. SynthID is an invisible watermark embedded into generated media. C2PA is signed Content Credentials metadata. Slop or Not checks both when the file exposes them.
Why can a SynthID check be inconclusive?
SynthID is embedded directly into the image, but repeated heavy edits, severe recompression, or too little image detail can weaken the signal. The image detector can still provide a probability when the watermark read is inconclusive.
What does the SynthID check add over the image model?
Evidence instead of a guess. Every image check here works the same way: it reads SynthID watermarks and C2PA or IPTC credentials, then scores the pixels with the image model when neither settles the file. A readable SynthID mark from a supported Google or OpenAI image is the strong case, because the mark lives in the pixels and survives screenshots and re-encodes. Without one, you get a probability from the image model, not a provenance answer.
Does Claude use SynthID?
For text, Anthropic says Claude's watermark is a version of SynthID-Text. For images, Claude uses C2PA Content Credentials, not image SynthID. Slop or Not reads image SynthID from supported OpenAI and Google/Gemini files and reads available C2PA credentials from Claude images.
Does this detect Nano Banana images?
Yes. Nano Banana is Google's name for the Gemini image model, and its output carries the invisible SynthID watermark like every other Gemini image. Slop or Not reads that watermark straight from the pixels and names Google as the provider, so a Nano Banana picture resolves the way a Gemini one does, screenshot included.
Is there a Grok or Sora image detector here?
Not a dedicated one, and there is a reason for that. Slop or Not reads SynthID from Google and OpenAI images only, so a Grok picture with no readable credential goes straight to the image model and comes back as a probability. Sora belongs to OpenAI, so a frame exported from it that still carries an OpenAI credential resolves through the C2PA read. Strip that and it is the image model again.
Can this SynthID detector check text?
Not for the watermark. The Text tab scores writing with an AI text detection model, in English only, and returns a probability. It does not read SynthID-Text watermarks. The SynthID read is the Photo tab.
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.
More from Numen
Two more apps that run on your device. Neither needs an account.
Private LLM
A Local AI chatbot for uncensored open-source models that runs offline on iPhone, iPad, and Mac. Private LLM uses OmniQuant and GPTQ quantization tuned per model, so its 3-bit models match the 4-bit models other apps ship.
Clean Links
Most tracking now hides behind link redirects. Clean Links unwraps short links and strips every tracking parameter from a URL before you share it, on iPhone, iPad, Mac, and the web.
Related checks
Compare Related Detector Workflows
Gemini AI Image Detector
Check whether a photo came from Nano Banana or Gemini image generation. Slop or Not reads Google SynthID and C2PA/IPTC metadata, then checks the pixels.
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.
C2PA Detector
Use this C2PA detector to check Content Credentials. Slop or Not reads signed provenance, then checks pixels when no provenance or watermark resolves the file.
