Many AI tools write a content credential into the files they produce, a built-in record of what made the file and what has happened to it since. When you upload a file, Frontify reads that record and shows it in your library. Developers can read the same information through the API.

Content credentials are based on C2PA, an open industry standard backed by Adobe, Microsoft, Google and others. Frontify builds on C2PA and doesn’t use a proprietary format, so credentials from tools that follow the standard are read the same way.

Frontify checks each file for a content credential when you upload it and reads it automatically. You don’t need to tag anything.

Frontify reads credentials in these file types:

The following file types are not supported:

Open any asset in a library. In the sidebar, scroll below the metadata section to find File origin.

Everything comes from the content credential in the file, except Certificate status, which Frontify works out by checking the credential. Content origin and Cert. generator are what the tool wrote into the file. Frontify doesn’t check that they’re true.

The file origin shows these fields:

If a file has no credential, the sidebar shows only two fields: Credentials as Missing and Content origin as No AI declared.

You can filter the backend library by content origin and certificate status, for example to find every file that declares AI generation before it’s used in a campaign.

Select the values to match. For Content origin, choose ‘is one of’ or ‘is not one of’, then pick from No AI declared, AI-assisted and AI-generated. For Certificate status, choose from Trusted, Unrecognized, Tampered, Missing.

For example, set Content origin to is one of AI-generated and AI-assisted to see every file that declares AI involvement.

The No AI declared filter shows files whose credential doesn’t mention AI, and files with no credential at all. It doesn’t mean those files were made without AI, because a credential can be missing or removed.

If you want to search by certificate status, to see which assets are Trusted, Unrecognized, Tampered or Missing, you can use the filter to quickly identify these in bulk.

Filters work in the backend library. They aren’t available on library pages. The same two filters are available through the API.

Anyone viewing an asset on a published library page sees its content origin and certificate status, in a simpler view than the sidebar. The same applies to library assets embedded in guidelines pages. There’s no setting to turn on, and it applies whether the library is public or restricted to certain people.

Downloading an original file keeps its content credential. If Frontify resizes, crops or converts a file, the credential isn’t carried over. Support for these versions is planned, with no date yet.

What is C2PA, and why does Frontify use it?
C2PA stands for the Coalition for Content Provenance and Authenticity. It’s an open industry standard for adding a verifiable record to a file, such as which tool made it. That record is the content credential. Because the standard is open, credentials from different tools follow the same format, so Frontify reads them without extra steps and doesn’t use a proprietary format.

What are content credentials useful for?
They show where a file came from, such as which tool made it and what the file says about AI involvement. In Frontify, your team can check that in the sidebar, filter the library by it before a campaign goes live, and show it to people viewing your published library pages. Developers can also read it through the API for reporting. How and when to disclose AI use stays with your team.

Who can see content origin and certificate status?
Anyone viewing an asset on a published library page sees its content origin and certificate status. This includes external viewers and applies whether the library is public or restricted to certain people. The same applies to library assets embedded in guidelines pages. There’s no setting to turn on. Please note: Content origin and Certificate status are not shown in public share links.

Can I read this through the API?
Yes. Developers and integrations can read the same file origin information through the public API, and the content origin and certificate status filters are also available there.

Does Frontify add content credentials to my files?
No. Frontify reads credentials that are already in a file. It doesn’t create, sign, or certify them. This includes files made in an AI app inside Frontify. If the app adds a credential, Frontify shows it. If it doesn’t, the file shows as Missing, with No AI declared.

What does “No AI declared” mean?
The credential in the file doesn’t claim any AI involvement. It doesn’t prove a person made the file, because the standard gives no guarantee of that. A file with no credential at all also shows No AI declared, alongside a Credentials field marked Missing.

Why does an AI-generated file show No AI declared?
Frontify shows what an asset’s content credential says. It doesn’t look at the image to guess if AI was used. Some AI tools don’t declare AI in their credential, or they sign it in a way Frontify can’t verify yet. In that case the asset shows No AI declared. This doesn’t mean a person made the asset, and it doesn’t prove AI wasn’t involved. If you know an asset was made with AI, your team’s own disclosure process should still cover it.

Why does my file show no credential?
Frontify can only show a credential if the file carries one. Not every AI tool adds one, and a credential can be lost when a file is screenshotted or re-encoded. Files in unsupported formats aren’t read. You can’t add a credential or mark an asset as AI-generated manually in Frontify.

What does “Unrecognized” mean?
Unrecognized means the file is signed and its content wasn’t changed after signing, but the signer isn’t on the C2PA or Frontify trust list, so Frontify can’t confirm who signed it. This can happen when a signer is new, for example a recently launched AI tool that hasn't been added to the list yet.

What does “Tampered” mean?
Tampered means the signature doesn’t match the file. The content was changed after it was signed.

Does the credential stay when I download a file?
Yes, for the original file you uploaded. Resized and converted versions don’t carry the credential. Support for these is planned, with no date yet.

Can I add a visible AI label to my assets?
Frontify doesn’t add visible labels onto assets. Origin information appears in the metadata within the sidebar and on published library pages, but nothing is added to the file itself.

Do my existing assets get processed?
Frontify reads credentials when a file is uploaded. A one-time scan of assets already in your library is available on request. It doesn’t run automatically. Request this service via your customer success manager.

Does this make my organization compliant with the EU AI Act?
Frontify detects and shows the content credentials that files already carry. Your team is responsible for how AI-generated assets are used and presented in published contexts. If an asset could mislead your audience, you may need to disclose AI use in that context. Under the EU AI Act, the duty to mark AI-generated output in a machine-readable way sits with the provider of the AI system that generates it.

Not seeing what you expect, or want more detail on how content credentials work? Contact your customer success manager.