DALL·E detector
DALL·E is one of the easier generators to identify, because OpenAI attaches C2PA Content Credentials to its output. That is a structured manifest embedded in the file stating which tool created it — far stronger evidence than any classifier can offer.
This tool reads that manifest directly. When it is present and names OpenAI, the answer is close to definitive. When the image has been re-saved or screenshotted, the manifest is gone and the check falls back to statistical detection with correspondingly lower confidence.
Drag a file here, or
JPEG, PNG, WebP, AVIF, PDF or .txt · up to 10 MB · you can also paste a screenshot
By uploading you agree to our Terms and Privacy Policy. Your file, IP address, approximate location and device details are stored.
What this checks
- C2PA manifests embedded in JPEG APP11 segments, PNG caBX chunks or WebP C2PA chunks
- The claim generator string, which names the creating tool
- IPTC trainedAlgorithmicMedia assertions inside the manifest
- Fallback image classifiers when no manifest survives
Frequently asked questions
Do all DALL·E images have Content Credentials?
Images produced through OpenAI's own surfaces generally do. Anything that has been screenshotted, re-encoded or passed through a platform that strips metadata generally does not.
Do you verify the manifest's signature?
No, and this matters. We detect the manifest and read the tool it names, but we do not check the cryptographic signature, so a hand-crafted manifest would be reported at face value. The result describes this as what the file declares about itself rather than as proof.
Can Content Credentials be removed?
Yes, trivially — a screenshot removes them. Their absence therefore says almost nothing, while their presence says a great deal. Evidence here is one-directional.