If JS a problem for you, try Kantu. It works with screenshots and uses OCR for scraping. The beauty is that it works with any kind of site. But clearly, the speed can not match a node.js or perl based scraper (mechanize etc), so it is not suitable for high volumes.