PurgeMeta

How to Remove C2PA Metadata & Prompts from ChatGPT Images (2026)

Tran Chi HuuJuly 28, 20264 min readChatGPT · C2PA · DALL-E 3 · Metadata Privacy

Why ChatGPT Embeds Metadata in Images

When you generate an image using OpenAI ChatGPT or DALL-E 3, digital provenance data is embedded directly into the file. This includes C2PA Content Credentials manifests, digital signatures, and the exact text prompt you used to generate the image.

What Information is Exposed in ChatGPT Images?

  • C2PA Provenance Manifest: Flags your image across social media platforms as AI-generated.
  • Generation Prompt: Exposes the original text prompt embedded in the PNG/JPEG headers.
  • Model Version & Seed: Discloses OpenAI model parameters.

Step-by-Step: How to Strip ChatGPT Metadata Safely

Using PurgeMeta, you can instantly sanitize your ChatGPT images without losing visual quality:

  1. Drag and drop your ChatGPT PNG or JPEG image into PurgeMeta Sanitizer.
  2. Leave Output format on Original for a lossless clean, or pick JPEG/WebP and a max size to convert and shrink at the same time.
  3. Click Save (Pro users can clean up to 200 images and download them all as one ZIP) to download a clean, metadata-free version.
All guides