How does Facebook get the images so quickly when posting a URL (Share a Link)?

To piggy back off of , one option you have is . I personally have used it on some of my larger scale apps and it helped a ton. Some reasons: oEmbed is simple, and though it's an open standard, there isn't a standard URL path companies use. In other words, you may find yourself having a lot of "ifs" in your code trying to determine what domain a media URL is pointing at, and then construct a path based on that finding ( http://fooA/services/oembed vs http://fooB/oembed/2 ). In fact, some companies even have different oEmbed paths for media types like /picture/oembed, /video/oembed, and /audio/oembed, so you'll have to account for this and provide wild cards as necessary ( /*/oembed ) If you want your photo extractor to work all the time, you're going to have to do some tedious research in accounting for all the different companies/services a person may take a URL from. To be fair, if you cover the top few players such as , , or , you'll be covering the 90% case. ( there are more than just those 3, and I'm not even sure you care about video for your use case ) I believe some media companies would actually rate-limit/CAPTCHA your service if it requested too many oEmbed extractions, in a specific timeframe. If this is still the case, you'll want to account for that as well. Embedly's customer service is top notch, and you get analytics for things like # of times a media came from YouTube and how many images were shared for any given day. Again, it really depends on how soon you need it, what media you're trying to detect, what services you want it to work with, and how much money you're willing to spend ( Embedly is free up to a certain point ).

Here's one way: Fetch per usual for the first share, then cache the result for subsequent shares (I can't say for sure but I would think that many links are shared by more than one person).