What is C2PA? It is an open standard — Coalition for Content Provenance and Authenticity — that embeds cryptographically signed content credentials into digital files. If you create or edit images with Adobe, OpenAI, Microsoft, or many AI tools, your exports often carry C2PA metadata. That is a common reason Instagram, Facebook, Pinterest, and TikTok show AI Info or Made with AI labels.
This page answers what is C2PA, what C2PA metadata looks like inside a file, and why platforms read it — plus where to remove it before upload.
Quick actions:
- Inspect your file: AI metadata checker
- Strip C2PA before upload: AI metadata remover
- Instagram-specific workflow: Remove AI Info from Instagram
What is C2PA? (short answer)
C2PA defines how to store a signed manifest inside image and video files. The manifest can record:
- Creation history — Which tools and services were used
- Modifications — Edits, AI generations, and transformations
- Assertions — Structured claims (e.g. AI-generated or human-created)
The goal is transparency: software that understands C2PA can read the manifest. For AI-assisted exports, assertions often state that AI was involved — and social platforms use that as an automated labeling signal.
What is C2PA metadata?
C2PA metadata refers to the manifest and its container blocks inside the file — not the visible pixels. In practice you may see:
- JUMBF content-credential boxes in JPEG or PNG
- APP11 or similar segments in JPEG
- XMP wrappers that reference C2PA manifests
- Container-level C2PA in MP4 and some video exports
Classic EXIF GPS cleaners often leave C2PA intact. If your post still shows AI Info after “removing EXIF”, check specifically for C2PA metadata with the metadata checker.
Who embeds C2PA metadata?
Common sources when you export or save:
| Tool / pipeline | Typical C2PA signal |
|---|---|
| Adobe Photoshop (Generative Fill, Neural Filters) | Partial-edit C2PA assertions |
| Adobe Firefly | Full C2PA on generated images |
| ChatGPT / DALL·E | OpenAI C2PA content credentials |
| Microsoft Designer / Copilot | C2PA on generated stills |
| Content Authenticity Initiative (CAI) cameras & apps | Capture or edit provenance |
The manifest is invisible in the image viewer but readable by upload pipelines that scan for provenance.
Why platforms use C2PA for AI labels
Instead of guessing from pixels alone, many platforms:
- Read the C2PA manifest on upload
- Check for AI-related assertions
- Apply labels such as AI Info or Made with AI
Strip C2PA metadata before upload and that structured signal is gone — when the label was metadata-driven. Pixel watermarks and separate classifiers are a different problem; see our disclaimer.
Related platform guides:
C2PA vs XMP vs EXIF
| Type | What it is | AI label role |
|---|---|---|
| C2PA | Signed content credentials / manifest | Strong automated signal on Meta, TikTok, Pinterest |
| XMP | Extensible metadata (prompts, seeds, model names) | Common on Midjourney, SD, Firefly exports |
| EXIF | Camera/software fields | Can name AI tools in Software tag |
A complete pre-upload workflow strips C2PA + XMP + relevant EXIF/IPTC AI fields. Our AI metadata remover targets all three in the browser.
C2PA manifest structure (technical overview)
A typical manifest includes:
- Claim generator — e.g. "Adobe Photoshop 25.0" or "OpenAI"
- Assertions — such as
c2pa.actionslistingc2pa.ai.generatedor partial edits - Signatures — cryptographic signatures for tamper detection
Platforms search assertions for AI involvement. That is why what is C2PA matters for creators: it is often the machine-readable reason a label appeared.
How to remove C2PA metadata
To remove C2PA metadata before uploading:
- Export your final JPG, PNG, or WebP
- Open the AI metadata remover or checker first
- Strip C2PA manifests and XMP AI tags
- Download the cleaned file and upload to your platform
Alternatives (less reliable):
- Re-export with metadata disabled in your editor — results vary by app version
- Screenshot — new file without manifest, possible quality loss
Bundled workflow (EXIF + C2PA + XMP): Strip metadata from images online.
Video / MP4:
- C2PA in videos — what it is and how to strip it
- Remove AI metadata from MP4 before uploading
- Does Instagram label AI-generated videos?
FAQ quick reference
- What is C2PA? → Open provenance standard; signed manifest in the file
- What is C2PA metadata? → The manifest + JUMBF blocks platforms scan on upload
- Does removing C2PA change pixels? → No, only invisible metadata
- Is EXIF removal enough? → Usually no — target C2PA explicitly
Conclusion
Understanding what is C2PA and what C2PA metadata contains helps you interpret AI Info labels. When labels are file-driven, removing the manifest before upload is the practical fix. For a one-click workflow, use the free metadata remover — then cross-post the same cleaned file to Instagram, TikTok, or Pinterest if needed.
Remove C2PA metadata before social upload
Strip the C2PA manifest and related XMP fields, then upload the cleaned image.
- Export your image — Save the file from Photoshop, Firefly, ChatGPT, or another C2PA-aware tool.
- Open a C2PA-capable remover — Use a browser tool that lists C2PA blocks before removal.
- Remove C2PA and XMP AI tags — Download the cleaned file after processing finishes locally.
- Upload to your platform — Post the cleaned asset so automated metadata checks no longer see the manifest.
Frequently Asked Questions
What is C2PA?
C2PA (Coalition for Content Provenance and Authenticity) is an open standard for embedding signed content credentials in images and video. The manifest records which tools created or edited a file and can include AI-related assertions that platforms use for disclosure labels.
Can C2PA provenance metadata be removed?
Yes. C2PA provenance metadata can be removed from JPG, PNG, and WebP in your browser without changing pixels — strip the manifest and JUMBF content credentials with a metadata remover, then upload the cleaned file before posting to Instagram, Pinterest, or TikTok.
What is C2PA metadata?
C2PA metadata is the signed manifest and related JUMBF blocks stored inside a JPEG, PNG, or video container. It is separate from classic EXIF GPS fields and from many XMP tags — platforms scan it for AI provenance signals at upload.
What does C2PA stand for?
Coalition for Content Provenance and Authenticity. Members include Adobe, Microsoft, and other technology and media companies building interoperable content credentials.
Why does C2PA trigger Made with AI labels?
Platforms read the manifest and look for AI-related assertions. When they find them, they can show AI Info or Made with AI disclosures automatically on Instagram, Facebook, Pinterest, and other uploads.
Is deleting EXIF the same as removing C2PA?
No. C2PA lives in its own structured blocks separate from classic EXIF GPS fields. You need a remover that explicitly targets C2PA manifests and JUMBF content credentials.
Can I remove C2PA without losing pixels?
Yes. Metadata-only removal keeps the same pixels while deleting the manifest and related AI tags. Use a browser-based remover that lists C2PA blocks before and after cleaning.