An option has been added to select how scraping is performed; It has four options: auto (default), native cURL calls, XenForo's client (Guzzle), or disabled. The default value is fine for the overwhelming majority of users so you don't have to care about it unless you use a special configuration for your network, in which case you should select the default XenForo client. Disabling scraping may prevent some content from being embedded.
The current default is to use native cURL calls because XenForo's client (Guzzle) does not seem support HTTP/2 at the moment. Future versions will eventually switch to use Guzzle, and the option will be removed.
In addition, defunct site Gfycat has been removed from the add-on and Facebook has been updated to support more video URLs.
An option has been added that will display a link to the original URL used to embed content. This is only available when the add-on is configured to use a [URL] BBCode instead of the [MEDIA] BBCode. The option is enabled by default buy only applies to content embedded via a [URL] BBCode.
This feature is accessible through the "Insert Media" dialog when posting a message. It adds two options (unchecked by default) that allows the user to look for embeddable URLs if the one they submit does not match any active media site. For example, if a user tries to submit a t.co (Twitter's URL shortener) link to a YouTube video it will not be recognized by default, but if the user checks "Use the canonical URL" then it will be recognized as a YouTube URL.
You can enable or disable this feature in your admin panel. When you install the add-on for the first time or you upgrade to this version, the status of this feature will match your preference for unfurling. If you have unfurling enabled, then this feature will be enabled. Otherwise, it will be disabled and you will have to manually enable it in the add-on's options.
This feature will evolve based on organic usage and your feedback.