Codex is our metadata repository / API. We use it to very efficiently access metadata for recommendation computation and display. Metadata for multiple type of content (Programs, audio and video episodes, articles and more) can easily be stored and updated. Additional data, like transcripts or extracted topics can also be stored.
This is a quite handy tool if your organisation is dealing with multiple systems and APIs, when API should not be accessed too frequently, or when data related to a given resource has to be fetched and consolidated from several sources. Processing of audio to automatically create transcript is quite commonly done outside of main system, and stored on the side. Codex conveniently enables their storage and usage in recommendation, increasing the discoverability and value of your media archives.