The oEmbed standard is great for exposing embeddable content in way that, most of the time, the average user never has to think about it -- it just works. However, with more complex data, including multiple content objects or underlying metadata formats, something less opinionated is necessary. This is the niche unAPI plays a simple, but powerful, role. Like oEmbed and other standards, unAPI defines a service endpoint with a handful of basic operations and discovery mechanism, and does so in a plain and obvious manner that makes it easy for tech-inclined folk to work with.
For an archives project at work, we implemented unAPI as a simple way to segment the page content and expose underlying metadata formats, in order to offer our partners a quick (and content agnostic) way to pull elements into existing tools. The API endpoint is exposed within the page header:
These formats could be used to share any type of data -- different flavors of metadata, content outside the application context. As with oEmbed, this provides a basic way to provide federated and aggregated data within a common framework. In addition to being a convenient service to share content among applications, there is also support for unAPI within the Zotero citation management tool.
What could this look like in order to access NPR.org content? Instead of a rich API, by simply feeding a story url into the unAPI service, an application could retrieve the different content elements -- a text/html or text/plain representation of the story, the audio/mp3 from the broadcast, an image/jpeg feature image, and perhaps an application/xml+rss feed of the series, comments, or category. It lacks the power behind the stand-alone API, but provides data in a form that is a little easier to craft new ways of highlighting this content on station websites.