[catch23]

This reminds me of a blog post I wrote a year ago about proxy technology:

http://www.deanmao.com/2012/08/28/modify-a-site-you-dont-own...

I did it the same way -- parse & rewrite js code, and have a server that stores the same cookies that your browser would. The browser js code also contains the same js rewriter so that it can handle stuff like eval(). It also had to do lots of header manipulation since there are lots of headers that have domain-name based security, like CORS for example. The one I wrote works great with js-heavy sites like gmail or facebook.