Just wanted to point out that pulling data from a set of sites with structured page layouts like Yummly does is relatively easy from a technical perspective, and many companies have done this. Say Mmm has taken this a step further and developed a solution a year and a half ago to tackle the much bigger challenge of parsing ingredient and other recipe information from any web page or even plain text. Since then, the parsing logic has been continually improved as users have added hundreds of thousands of recipes on our site and through our Evernote integration to automatically convert ingredients into shopping lists with nutritional estimates.