oEmbed is a brand-new spec, announced by Cal Henderson (Flickr), Mike Malone, Leah Culver (Pownce), and Richard Crowley (OpenDNS).
oEmbed is an open format for standardizing how images, videos, links, or other media get embedded, and for sidestepping media providers’ APIs (or the screen scraping you’d need when they don’t have one). When a user posts a link that points straight at a specific media resource, it lets a site embed that media quickly and painlessly. Here’s how it works: it converts the link accordingly — that is, it turns the image or video into XML or JSON that tells you how to embed the media.
“oEmbed is a format that allows an embedded representation of a URL to live on third-party sites. When a user posts a link to a resource, a simple API lets a website display embedded content (such as photos or videos) without having to parse the resource directly,” the authors say on the oEmbed page.
Example: http://flickr.com/services/oembed?url=http://flickr.com/photos/bees/2362225867/ returns:
Early adopters of the oEmbed spec include Flickr, Viddler, Pownce, Qik, and Revision3. Viddler has already put together a form for testing the web service, so you can see it in action.
At its core, oEmbed makes it easier to get at images and videos using plain URLs. Standardizing how a resource gets embedded means, for instance, that a community site can let users easily embed media from any oEmbed-enabled site by simply pasting in the URL of a picture or video — and the “client” site won’t need to wire up the APIs of N different media providers.
This is a really great idea! Just think — we won’t have to deal with each type of displayed media one by one at all.