For now it's only using the search results from Bing (it doesn't read the whole pages), we will have a scraper very soon