AI image detector by generator
ChatGPT image detector: what the download proves
Images generated in ChatGPT carry C2PA Content Credentials signed by OpenAI, so isGenAI can show a validated creation record for the original download; a screenshot carries nothing to read.
test-signed.pngPNG file · controlled example
- Signature + IHDR
- iTXt XML:com.adobe.xmp
- tEXt parameters
- tEXt prompt / workflow
- tEXt invokeai_metadata
- eXIf
- caBX C2PA manifestrecord
- IDAT pixels
- IEND
Content Credentials found
A signed manifest describes how the file was made.
Signer, recording tool, actions and the signature check result are reported separately, so you can weigh each one.
Open the Content Credentials checker →What ChatGPT writes into generated images
OpenAI’s help centre states that images generated with ChatGPT, Codex and the OpenAI API include both C2PA metadata and a SynthID watermark, and that ChatGPT images include an additional reference to their ChatGPT origin. The C2PA manifest is a signed record of the generating tool and action; the watermark is an invisible pixel-level signal intended to persist when the metadata is removed.
OpenAI also documents the file-size effect of the manifest and warns that most social media platforms remove the metadata and that a screenshot removes it. Some OpenAI community reports describe exports that arrived without a credential, so always verify the file you actually hold.
What isGenAI shows for a ChatGPT image
Upload the PNG or WebP as downloaded and the report lists the signed creation record: signature validity, signer trust, claim generator and the AI generation action. The source type trainedAlgorithmicMedia inside the manifest is what turns it into strong, direct evidence of AI use.
isGenAI does not read SynthID. If the manifest is gone, the report says no file evidence was found and offers the matching-copy check against records that still carried the credential.
Copies, edits and reposts
Editing a ChatGPT image in an app that does not understand Content Credentials breaks the hash binding, so the manifest may validate as tampered or be dropped entirely. The checker reports a broken binding honestly instead of hiding it.
For an image seen on Instagram, X or WhatsApp, the realistic path is to get the original export or the first upload. The platform pages on this site explain what each service strips.
How to check an image from this generator
- Get the original download, not a screenshot or a copy saved from a chat or social feed.
- Upload it to the free AI image detector. Files up to 20 MiB in PNG, JPEG, WebP, AVIF or TIFF are read on the server and are private by default.
- Open “How we checked” and read each finding with its source field: signed record, declared source type, saved prompt, AI statement or matching copy.
- Treat a clean result as “no evidence found”, never as proof that a person made the picture.
Frequently asked questions
Does ChatGPT add metadata to generated images?
Yes. OpenAI documents C2PA Content Credentials and a SynthID watermark in images generated with ChatGPT, Codex and the API, plus a reference to the ChatGPT origin, as of October 2026.
Can isGenAI detect a ChatGPT image after a screenshot?
No. A screenshot has no C2PA record and isGenAI does not read pixel watermarks. The report states that no evidence was found and suggests obtaining the original file.
What is the difference between the C2PA record and the watermark?
The C2PA record is signed metadata inside the file that isGenAI validates. The SynthID watermark is an invisible pattern in the pixels that OpenAI and Google tools can check; isGenAI reports only the file record.
Why would a ChatGPT download show no credential?
Re-saving, converting, editing in a non-C2PA app or a phased rollout can leave a file without the manifest. OpenAI states that an image without the metadata may or may not have come from its tools.
Sources
Vendor behaviour changes. These pages describe what the vendor documents and what isGenAI reads from a file; they do not test every release.