Arbor Media
Explore ArborArbor Signum demo
Demo · fictitious content
← All demos
In pilotBuilt and being tested in pilots.

Arbor Signum · Content Credentials (C2PA)

Can you trust what you see?

Every video carries its own proof of origin. Click the Content Credentials mark in the player to see who published it, where it came from and whether it has been changed since.

Valid credentials Invalid credentials No credentials

Media centre

Official video statements

Video statements published by a fictitious parliament on Arbor WebTV. Content Credentials are shown where provenance information has been attached.

0:00 / 0:00
Example video with valid credentials
Content Credentials valid

Digital trust statement

Speech by MP Julien Marchand at today's parliamentary session on digital trust.

22 Jul 2026
Arbor Signum
0:00 / 0:00
Example video with invalid credentials
Content Credentials invalid

Budget fund allocations

Court of Justice of the European Union - Recording where validation does not match the original and causes a failed Content Credentials signal

22 Jul 2026
Does not match
0:00 / 0:00
Example video without credentials
No Content Credentials

Protecting citizens

An impassioned speech by MP Mara Elsenbrück on protecting citizens.

22 Jul 2026
Unknown

Story · The altered clip

When someone alters your clip

Someone downloads an official clip, replaces the speaker’s face and posts it again. The Content Credentials are gone, but the invisible watermark is not. See how a communications officer uses the Arbor Signum Verify Portal to find the original and see exactly what was changed.

Watch the story

How it works

How Arbor adds Content Credentials and watermarks

Every recording gets an invisible watermark and a signed C2PA manifest. When clips are combined, the new video keeps a record of where each part came from.

1When a session is recorded

How Content Credentials and watermarks are added to a recording Arbor Record captures the session. An invisible layer 1 watermark is embedded in the video. A C2PA manifest is signed with Arbor Signum. The video is published and viewers check the Content Credentials mark in the player. Record Arbor Record captures the session Watermark · layer 1 Invisible watermark embedded in the video Sign C2PA manifest signed with Arbor Signum Publish and verify Viewers check the Content Credentials mark in the player

2When clips are combined with Clip & Share

How Content Credentials are kept when clips are combined Clip A and clip B come from recordings with a layer 1 watermark and their own C2PA manifests. Clip C is a plain video without credentials. Clip & Share combines them into one rendered clip, which gets a layer 2 watermark and a new C2PA manifest signed with Arbor Signum. The new manifest lists manifest A, manifest B and clip C as its ingredients. Clip Afrom recording A Watermark · layer 1 C2PA manifest A Clip Bfrom recording B Watermark · layer 1 C2PA manifest B Clip Cplain video No watermark No C2PA manifest Clip & Share Select and arrange clips into one video Rendered clip One new video Watermark · layer 2 added on top of the clip New C2PA manifest signed with Arbor Signum Clip manifest Records every source INGREDIENTS Manifest Aclip A, signed Manifest Bclip B, signed Clip Cplain, no credentials

Clips with and without credentials can be combined. The new manifest points to the original manifests and notes any content that had none.

C2PA signing and the Content Credentials mark in the player are part of Arbor Signum Publication. Invisible watermarking is part of Arbor Signum Protection.

FAQ

Frequently asked questions

Click a question to see the answer.

Does Arbor Signum protect the audio track within a video?

The audio is protected by the signature; the watermark currently covers the video. Audio and video are hashed and signed together as one audiovisual package, so any change to the audio is detectable. The invisible watermark is applied to the video only for now.

Audio watermarking is on our roadmap. If it becomes a formal requirement for you, we can revisit our delivery priorities together.

Will I see the Content Credentials logo on live sessions?

Not yet. Live sessions do not yet carry C2PA information, so the Content Credentials logo does not appear during a live broadcast. Live support is on our roadmap.

Does watermarking affect latency, video or audio quality, or encoder load?

Watermarking adds a small delay of a few milliseconds. We have not yet quantified the effect on image quality, audio quality or encoder load. We will measure all three during the pilot, so you can decide on figures from your own environment.

What happens if the signing environment is temporarily unavailable?

Watermarking and signing are separate processes. Watermarking happens during encoding and stays in the encoded output. Signing is part of publication: if signing fails, publication stops as well, and that triggers an investigation.

Where is the watermark added?

The Arbor Encoder adds a Layer 1 watermark during encoding. The clipping tool can add a Layer 2 watermark to clips. Layer 1 watermarking also works for content that arrives without an existing watermark. Recovery is designed to follow the same principle, although full C2PA support in that process is still in development.

Which downstream operations can remove or invalidate Content Credentials?

Tools, encoders and editing processes that do not support C2PA can remove the credentials or make them invalid. That is by design: a modification outside a C2PA-aware workflow breaks the chain of provenance, so anyone checking the recording can see that it has changed.

C2PA-aware tools read the existing provenance information, keep it and add their own editing steps.

How robust is the watermark against editing?

You decide how robust it needs to be. The watermark can survive operations such as cropping and re-encoding, to a configurable degree. Greater robustness can make the watermark more visible, so we tune the setting with you and validate it during the pilot.

Does Arbor Signum change codecs or file formats?

No. Apart from embedding the watermark, Arbor Signum leaves the codec of the Arbor Encoder output unchanged. Format and codec compatibility stays as it is today, determined by your current encoding and distribution chain.

Can partners, broadcasters and social platforms keep Content Credentials intact?

Yes. C2PA information is openly readable. Partners can keep and extend the credentials when their tools support the public C2PA specification. Processing that does not support C2PA may remove or invalidate them.

We therefore recommend agreeing with your chain partners that they use compatible workflows and retain provenance metadata.

Can Arbor Signum later support provenance from camera and microphone onwards?

Yes. Keeping and extending credentials from upstream devices has been the intended direction from the start. There is currently no confirmed way to carry C2PA information from the camera through the SMPTE ST 2110 contribution path into this workflow. We will implement it once a clear, interoperable method is available.

Who owns the signing certificate and key?

The certificate is issued to your organisation, which applies for it and owns it. Arbor needs controlled access to use the signing key for the service. The key is stored in Azure Key Vault Premium, protected by a hardware security module.

Is use of the signing key logged?

Yes. Key usage is recorded through Azure audit logging.

Can you revoke the certificate?

Yes. Because the certificate is yours, you can ask the certificate provider, such as SSL.com, to revoke it.

Can I use Arbor Signum as an add-on to another streaming platform?

Arbor Signum is an add-on to an Arbor product. It is not available on its own for other platforms.

Can metadata be included in Content Credentials?

Yes, in principle. Textual metadata such as speakers, agenda items and timecodes can be included in the C2PA manifest, so the context of the session travels with the recording. Together we decide which metadata is useful and proportionate from a privacy perspective.

Which personal data will be included, and how will privacy be assessed?

You decide which personal data, if any, goes into the manifest.