Blog
What Is C2PA Metadata in ChatGPT Images? (Plain-English Guide)
2026-05-12 ยท 7 min
C2PA is the invisible cryptographic stamp OpenAI bakes into every ChatGPT image. Here's exactly what it is, what it contains, and how to remove it.
Every image ChatGPT generates carries an invisible passport. It is not a pixel you can see and it is not metadata you can read in your phone's gallery โ it is a cryptographically signed block called C2PA Content Credentials, embedded in a container called JUMBF.
What does C2PA stand for?
C2PA stands for the Coalition for Content Provenance and Authenticity. It is a standards body founded by Adobe, Microsoft, the BBC, Intel, Truepic and others. Their goal: give every digital file a tamper-evident chain of custody so the world can tell AI-generated content from camera-captured content.
What's actually inside the watermark?
- The generating model (e.g.
gpt-image-1,dall-e-3) - The exact timestamp of generation
- A cryptographic signature from OpenAI's signing certificate
- Software details and JUMBF box hierarchy
- Hash of the image bytes
Anyone with a C2PA viewer โ or even exiftool โ can read all of that in seconds.
How to remove it
Open our free ChatGPT image watermark remover, drop your image in, and download the cleaned copy. Pixels are byte-identical to the original. Metadata is gone.
Ready to remove your ChatGPT watermark?
100% free ยท no login ยท images never leave your browser
Open the Free Tool โ