I hand-rolled a wells fargo scraper recently using the haskell webdriver package and Selenium. It was working fairly well until they started requiring a captcha - something I've never seen as a human user. I'm not quite sure what triggered that or how to get around it.
