Video in Every Experience, Governed From One Library

What you can do together

• Place one embed in a component and have every page built from it carry the player, without a video file ever entering the AEM repository.
• Name the domains a video may render on, and refuse the rest, with signed URLs that expire so an address cannot be passed around.
• Keep an unpublished or restricted video governed in Nexus while its page publishes normally, the player simply absent for a viewer without access.
• Let an agent work against AEM content through Adobe's own server, so pages referencing a retired video can be found rather than hunted for.
• Read engagement per surface, so the same video's performance differs by the site it appeared on rather than being one aggregate number.

How it connects

Rendering is the JavaScript widget. An author places an embed in a component or an experience fragment, and every page or template built from it carries the player without the media entering the AEM repository. The documented set covers video, audio, image, document, collection and playlist players, media library gallery and detail views, insights, transcription, maps, carousel and collage, quiz and survey players and a mashup filter, so a page can carry a single video or a browsable library view. This is the same path the shipped WordPress and SharePoint integrations run on.

Because one AEM instance usually serves many domains, the approved-domain list is the control that matters. Embedding is restricted to the domains you name, so each published site renders video and anything else is refused, and content URLs are signed and expire rather than remaining valid once shared. Each embed resolves against the viewer's identity, so an item restricted in Nexus renders without its video for a visitor who may not see it, on a public site and behind a login alike.

Adobe also runs a hosted MCP server for AEM as a Cloud Service, and AI Intelligence Hub connects to MCP servers as a client. The connection is configured once with a name, the server URL, the transport and the headers carrying its credential, and the tool list is fetched when it binds rather than coded in. What an agent can read or change in AEM is bounded by what that credential is granted there, which is the thing to scope at setup.