This is actually really cool! Are you just searching for content in the description, or do you scrape the link's contents too to get your curation of content?