EraseAILabelHome / What is C2PA
What Is C2PA, and What Are Content Credentials?
Content Credentials are a record attached to an image describing how it was made and edited. Here is how they work and what they can and can't tell you.
Last reviewed November 2026
The short version
C2PA stands for the Coalition for Content Provenance and Authenticity, a group formed in 2021 by companies including Adobe, Arm, BBC, Intel, Microsoft, and Truepic. It publishes an open technical standard for recording where digital content came from. Content Credentials is the consumer-friendly name for that information.
How it works
When a supporting camera or app creates or edits an image, it can add a signed record called a manifest. The manifest makes statements, known as assertions, such as which tool was used, what edits were made, and whether generative AI was involved.
The record is cryptographically signed and linked to the image data. If someone later changes the picture without updating the record, verification tools can detect that the two no longer match. In JPG files the data is usually stored in a container format called JUMBF.
Where you may encounter it
- Creative tools from Adobe, including Firefly.
- Images generated with some AI services, including tools from OpenAI.
- Certain professional cameras from makers such as Leica and Sony.
- Platforms that read these signals to show information about content.
What Content Credentials can and can't prove
A valid record shows who signed it and that the data hasn't changed since. It does not prove that a picture is true, fair, or complete. A misleading photo can carry valid credentials.
The absence of credentials proves nothing either. Most images don't have them, and the data is often lost when a file is screenshotted, re-saved, or compressed.
Metadata labels versus watermarks
| Content Credentials | Pixel watermark | |
|---|---|---|
| Stored in | The file's metadata | The image itself |
| Visible | Through a viewer | Usually not, for invisible marks |
| Example | C2PA manifest | Google's SynthID |
| Removed by re-saving | Often | Designed to survive common edits |
Our tool removes supported metadata only. It does not detect or remove pixel watermarks, and it isn't an AI detector.
How to check an image
The Metadata checker shows whether a supported file contains Content Credentials. It only detects that the block is there. To read the details and validate the signature, use the free Content Credentials Verify site.
Should you keep them?
Usually, yes. These records help people understand what they're looking at, and they can protect creators by showing the origin of their work. A person might reasonably remove them from a personal file for privacy, since a record can include names or editing history.
Removing a record doesn't change how an image was made. If an image is AI-generated, say so wherever disclosure is expected, and don't use metadata removal to pass it off as a real photo. See our Terms of Use.
Common questions
Is C2PA the same as an AI label?
No. C2PA is a general record of how content was made and edited, and it applies to real photos too. Some platforms use that information as one signal when they decide to show an AI label.
Can Content Credentials be faked?
The signature makes undetected changes difficult, but a valid signature only shows who signed the record. It does not prove that what the record claims is accurate.
Why does my image have no Content Credentials?
Most images never had them. They may also have been lost if the file was screenshotted, compressed, or saved by software that does not preserve them.
Check your own photo
See what your file carries. Nothing is uploaded.