Images made with ChatGPT carry C2PA Content Credentials. Here is what they contain and how to remove them without losing quality.
🔒 Your files never leave your device — processing happens in your browser.
What metadata does ChatGPT add?
Since 2024, OpenAI adds C2PA Content Credentials to images generated by ChatGPT and its image API. The manifest is signed by OpenAI and
states that the image was created by an AI model; Content Credentials viewers display it as “Issued by OpenAI”. Files downloaded from ChatGPT are also
named ChatGPT Image <date>.png, which reveals their origin on its own.
How to check a ChatGPT image
Drop it in the AI metadata checker: you will typically see a C2PA manifest block and the OpenAI signal.
How to remove it — step by step
- Drag your ChatGPT images into the box above (PNG or WebP, any number).
- Optional: choose Neutral names so the “ChatGPT Image…” file name is replaced.
- Download the clean files or the ZIP. The report confirms the C2PA manifest was removed.
The picture stays byte-for-byte identical: we only remove the manifest.
Good to know
OpenAI also keeps internal tools to recognise its images, and some platforms detect AI content visually. If you publish AI images where disclosure is required — ads, political content, marketplaces with AI rules — label them. Removing metadata is useful for clean deliverables, smaller files and privacy.
Frequently asked questions
Does ChatGPT add a visible watermark?
Generated images are delivered without a visible logo; the provenance information is stored in C2PA metadata.
Is the image quality reduced after cleaning?
No. Clean Image Metadata does not re-compress PNG or WebP — it only removes the metadata chunk.