Ah thats really interesting! How do you evaluate large scale cloud scraping services, since its operations are entirely hidden from you?
Personally I am looking into options in this area, are you planning to offer a cloud based version of this at some point/could you tell which existing ones are good if not?