AI tagging for footage on your own storage.

The shot nobody named

A folder name records a card, a date, or a project code. It rarely records the picture. The product on the table, the room, the hand in frame: those are how people remember a shot, and they are a poor match for a file called something like A001_C004. Asking one person who was in the room does not scale once the library does.

AI tagging is how that picture becomes something you can search, without first shipping the masters to a cloud service. The tags record what a pass over the frames decided the picture contains. Semantic search is how you ask for that picture later, rather than for the file name.

It runs on hardware you control

Media Nexus indexes a team's existing media storage in place. Approved volumes are read. Originals stay where they are. Media Nexus does not rename, move, or delete them. AI tagging and semantic search are part of that archive. Tagging runs on hardware the team controls, the same way transcription does. Nothing about the tags requires the library to leave the NAS or RAID.

Frame grabs give a picture of a moment. Tags sit with those frames. Search can use the file name, the tags, and, where someone spoke, the transcript. The working set is there so a person can find a moment before an editor opens the master. The master itself keeps its address.

This is not a SaaS MAM. A hosted catalog that needs an upload is a different shape, and it is compared in SaaS MAM versus keeping the files on your storage. Here the pass over the frames happens beside the files, on the system you own.

What you can ask

File-name search still works, and it is still the right tool when you know the card. Tagging is for the question you have when you do not.

• A product, a room, or a hand someone remembers, including when the file was never named for it.
• A shot in a long day of coverage, where the only clue is what was in the frame.
• A check before the edit. Playback confirms the hit, and the result points at the file already on the storage the edit uses.

Spoken lines are the other half. If the clue is a sentence rather than a picture, that search is search video by what was said. Most libraries need both, because some shots are silent and some lines are the whole point. Searching footage on a NAS is the page that holds the two together.

What the tags are not

A tag is not a new master, and it is not a reason to move the master. Proxies are the lighter files you play and hand off. They are generated beside the originals, and the originals stay untouched. What a proxy is, and why the archive makes one, is proxy files for footage on a NAS.

Tags also do not back the volume up. They describe what the approved volumes hold. They do not duplicate those volumes, and they do not tell you which shuttle left with an editor. Media Nexus is not a backup, and it does not move media offsite. Relay is a later build Vaultline scopes for a client after the archive, and drive tracking is one of its features. Full drive census, verified backup, and automation are later builds.

Scratch is still a human decision. The person who knows the library says which folders are the archive. Tagging follows that boundary. It does not invent a library out of every disk that happens to be mounted.

Who it is for

This is for in-house content teams whose footage already lives on storage they control. The pain is the shot only one person can find, and a library that gets harder as it grows. It is not aimed first at broadcasters, and it is not a plugin inside one editing application. Editors keep the tools they already cut in. Where the system is installed is on-prem media asset management.

The Free Build builds the archive on your real media before you decide whether to keep it. Build before you buy is the intake that configures that archive before any payment. The tags you see are tags of your footage, on your storage, not a demo library that had to be uploaded first.