True (well, it makes a canvas element! but canvas is bitmap-backed, and what the examples were doing was then extracting that bitmap to feed into a PDF file).
My points above were to avoid using PDFjs; if you use a server-side solution there is no need for it.
PDFjs basically takes javascript commands and constructs a PDF with them. In fact, the PDF format itself is just a list of drawing commands. So if you want to use it, you'd need to supply the commands from somewhere (which is basically making a HTML->PDF renderer!). Therefore there's no point to use this tool, as the other ones I mentioned directly take HTML and output PDF.
I'm not quite sure what you're asking here, but PDF.js is basically just PDF viewer made in javascript to bypass issues that have historically happened across different browsers to show PDFs. You can probably integrate this in Hype if you wanted.