About Wikidata

Wikidata is a free, open database of people, places, organizations and things, run by the Wikimedia Foundation. Every person in it has a permanent ID and a full set of names, titles and nicknames in many languages.

Moments Lab and Wikidata

Before, your team had to know how a person was spelled in your own list. A producer searching "King Charles" got nothing back if your Thesaurus said "Charles III," and every new joiner learned the house spelling the hard way. Now any person in your Thesaurus can be linked to their Wikidata entry, and every name that entry carries becomes a way to find them in your footage. Search "the King of England" and you land on King Charles III, in your own shows, in whatever language your editor thinks in.

How it works

Face recognition in Moments Lab is trained from your Thesaurus. You add a person, add a few photos of them, and the AI finds them across the catalog from then on. Wikidata sits on top of that. It is an option per person, not a replacement for your own list.

What goes out, and what comes back

Moments Lab looks up the Wikidata entry by its ID and pulls back the person's name, description and aliases, in every language that entry holds. You can also point a face training folder at a Wikidata ID instead of typing a name, and the person is created already named in all those languages. People with no Wikidata entry, and your own cast and contributors who will never have one, work exactly as before.

What your team gets

Every alias on that entry becomes a way to find the person in your footage. Your editor searches "the King of England" and lands on King Charles III, whatever your Thesaurus calls him and whatever language they searched in. The person stays one entry, so the same face across shows, seasons and spellings does not split into three.

‍