I wrote a node script[1] you could use for this.

[1] http://pastebin.com/raw.php?i=NgTeseN1

Here's a streaming version that runs through the top 1m Alexa sites. It looks like it gets throttled after a while though, but you can fix it with some trial and error by dialing down the concurrency in concurrent-map-stream and by introducing pauses.

https://gist.github.com/anonymous/4e34a10f1552dd8ede96

Cool idea to use the alexa top 1m.

Got this error though:

phantom stdout: TypeError: 'null' is not an object (evaluating 'element.value = text')

	  phantomjs://webpage.evaluate():3
	  phantomjs://webpage.evaluate():4
	  phantomjs://webpage.evaluate():4
	TypeError: 'null' is not an object (evaluating 'element.value = text')