How to Remove C2PA Content Credentials from Images
C2PA Content Credentials are the #1 reason platforms add "Made with AI" labels to your posts. StripShot removes C2PA data at the binary level, preventing automatic AI detection on Instagram, Facebook, and LinkedIn.
What are C2PA Content Credentials?
C2PA (Coalition for Content Provenance and Authenticity) is a standard created by Adobe, Microsoft, Intel, and others. It embeds invisible "Content Credentials" into image files that record how the image was created, including whether AI was used. This data is stored as JUMBF (JPEG Universal Metadata Box Format) boxes in JPEG files and as special chunks (caBX, caMs, caSt) in PNG files.
Which platforms read C2PA data?
Instagram & Facebook (Meta)
Meta reads C2PA Content Credentials and automatically adds a "Made with AI" label to posts. This label cannot be removed after posting. The only way to prevent it is to strip the C2PA data before uploading.
LinkedIn displays C2PA Content Credentials as a "CR" icon on images. Clicking it shows the full provenance chain, including AI generation details.
Pinterest uses a combination of C2PA metadata and AI detection to flag and suppress AI-generated pins in search results and recommendations.
How StripShot removes C2PA
Most metadata tools skip C2PA entirely because it uses a different embedding format than traditional EXIF/XMP metadata. StripShot specifically targets C2PA data at the binary level:
- 1JPEG files: Scans for APP11 markers (0xFFEB) containing JUMBF boxes. These are the containers for C2PA Content Credentials. StripShot removes the entire APP11 segment while preserving all image data.
- 2PNG files: Identifies and removes caBX (Content Authenticity Box), caMs (CA Metadata Store), and caSt (CA Store) chunks. These PNG-specific chunks carry the same C2PA credential data.
- 3Zero quality loss: Unlike tools that re-encode through a canvas element (converting everything to lossy JPEG), StripShot operates directly on the binary data. Your pixels stay identical.
Which AI tools embed C2PA?
| AI Tool | C2PA Embedded? | Details |
|---|---|---|
| DALL-E / ChatGPT | Yes | Full C2PA manifest with OpenAI as creator |
| Adobe Firefly | Yes | Adobe Content Authenticity Initiative credentials |
| Midjourney | No (EXIF only) | Uses EXIF Description/Comment fields instead |
| Stable Diffusion | No (tEXt/EXIF) | Parameters in PNG tEXt chunks and EXIF |
| Microsoft Designer | Yes | C2PA manifest marking AI generation |
| Adobe Photoshop (with AI features) | Yes | Even minor Generative Fill adds C2PA credentials |
Important: Adobe Photoshop users
If you use Adobe Photoshop's Generative Fill, Remove Background, or any AI-powered feature, Adobe automatically adds C2PA Content Credentials to your saved file. This happens even if the image was originally a real photograph. The "Made with AI" label can appear on photos you took yourself, just because you used an AI-powered editing tool.
Anti-fingerprint mode for extra protection
Some platforms use perceptual hashing to detect AI-generated images even after metadata removal. StripShot's optional anti-fingerprint mode applies sub-perceptual pixel modifications that change the image hash without any visible difference. This defeats hash-based matching systems that compare uploaded images against known AI-generated image databases.
Anti-fingerprint is optional and separate from metadata stripping. You can strip metadata (lossless, binary-level) without enabling anti-fingerprint, or enable both for maximum protection.
Free. No sign-up required. Files never leave your device.