What Is C2PA?
Quick Answer
C2PA (Coalition for Content Provenance and Authenticity) is an open standard for embedding signed provenance metadata in media files. It identifies images as AI-generated and specifies the creating tool. Instagram, TikTok, LinkedIn, and Pinterest read C2PA data to apply Made with AI labels automatically.
Where C2PA lives in your file
C2PA Content Credentials are stored in dedicated binary segments of the file, separate from the pixel data. The exact location depends on the file format.
| File Type | Location | Byte Marker |
|---|---|---|
| JPEG | APP11 marker segment (JUMBF container) | FF EB |
| PNG | caBX chunk | chunk type header |
| Video (MP4/MOV) | C2PA uuid atom | ISOBMFF container |
Inside a JPEG JUMBF block, the C2PA manifest contains content blocks: caBX (Content Authenticity Box), caMs (CA Manifest), and caSt (CA Store). The manifest includes a cryptographic signature that links the metadata to the file at the time of generation.
Which AI tools embed C2PA
Not all AI image generators embed C2PA. Tools built on Adobe, Microsoft, and Google infrastructure are the primary sources. Midjourney and Stable Diffusion use different metadata formats.
DALL-E 3 (via ChatGPT)
Full C2PA manifest
Adobe Firefly / Canva AI
Full C2PA manifest
Google Gemini image gen
Full C2PA manifest
Microsoft Designer
Full C2PA manifest
Midjourney
Uses EXIF, not C2PA
Stable Diffusion
Uses PNG tEXt chunks
How to remove C2PA
Because C2PA lives in specific binary segments of a file, it can be removed without touching the pixel data. Pixels are unchanged. File quality is unchanged. The only difference is the absence of the credential blocks that triggered the AI label.
Drop your image into StripShot
The tool reads the raw binary of your file. Nothing is uploaded to a server. Processing happens in your browser.
StripShot detects the C2PA blocks
For JPEGs: the scanner finds APP11 segments containing JUMBF C2PA manifests (byte sequence FF EB). For PNGs: caBX credential chunks. For video: C2PA uuid atoms.
Download the clean file
The APP11 segment (or caBX chunk, or uuid atom) is excised from the binary. The remaining file structure is valid. No re-encoding. Upload to Instagram, TikTok, or LinkedIn with no AI label trigger.
Frequently asked questions
What is C2PA?
C2PA stands for Coalition for Content Provenance and Authenticity. It is an open technical standard developed by Adobe, Microsoft, Intel, BBC, and others for embedding signed provenance metadata in media files. C2PA Content Credentials declare the origin of an image (AI-generated or camera-captured), identify the creating tool, record the timestamp, and include a cryptographic signature. Instagram, TikTok, Facebook, LinkedIn, and Pinterest read C2PA data to apply Made with AI labels automatically.
Where is C2PA stored in an image file?
In JPEG files, C2PA data lives in the APP11 marker segment, identified by byte sequence FF EB. The data is formatted as JUMBF (JPEG Universal Metadata Box Format) and contains content blocks including caBX (Content Authenticity Box), caMs (CA Manifest), and caSt (CA Store). In PNG files, C2PA lives in a dedicated caBX chunk. In video files (MP4/MOV), it is stored in C2PA uuid atoms in the ISOBMFF container.
Which AI tools embed C2PA?
DALL-E 3 (via ChatGPT), Adobe Firefly (and Canva AI), Google Gemini image generation, Adobe Photoshop (AI features), and Microsoft Designer all embed C2PA Content Credentials. Midjourney and Stable Diffusion do not embed C2PA but embed other forms of metadata (EXIF and PNG text chunks respectively).
Does Instagram use C2PA?
Yes. Instagram reads C2PA Content Credentials in uploaded images and videos. If a file contains a valid C2PA manifest identifying it as AI-generated, Instagram applies a Made with AI label that is visible to viewers. This process is automatic and does not require user disclosure.
Can C2PA be removed?
Yes. C2PA Content Credentials are stored in specific binary segments of image files and can be surgically removed without affecting the pixel data or image quality. StripShot removes C2PA from JPEG files by targeting the APP11 marker segment, from PNG files by removing the caBX chunk, and from video files by removing C2PA uuid atoms.
What is the difference between C2PA and a watermark?
C2PA is metadata stored in the file's binary structure, separate from the pixel data. It can be removed by editing the file's binary segments without altering the image visually. A watermark is embedded in the pixel data itself, either visibly (a logo overlay) or invisibly (steganographic pixel modifications). Pixel-level watermarks like Google SynthID survive metadata removal because they exist in the image content, not the metadata structure.
Is C2PA required by law?
Not in most jurisdictions yet. The EU AI Act (Article 50) requires AI-generated content to be labeled, but does not mandate C2PA specifically. C2PA is one technical implementation of that requirement. US law currently has no federal mandate. Social platforms have voluntarily adopted C2PA-based labeling ahead of regulatory requirements.
What does C2PA stand for?
C2PA stands for Coalition for Content Provenance and Authenticity. It is an industry consortium founded in 2021. Founding members include Adobe, Arm, BBC, Intel, Microsoft, and Truepic. The C2PA specification is maintained at c2pa.org.
Remove C2PA from your images
Binary-level removal. No re-encoding. No quality loss. Free for 3 images per day.