PhantomJS uses the same mechanism that wkhtmltopdf uses (headless webkit) and thus requires the 'server-side trickery' mentioned by the parent.

Also, the correct link is http://phantomjs.org/