Xpcom C++ : how to extract text from DOM elements

Hello ! With my embedded Mozilla web browser i would like to output a
text only representation of a loaded url. Because small is beautiful
and i just want to extract essential informations from pages, without
images, colors, fonts...
Is there xpcom interfaces to manage that ? Which interfaces/methods
gets the text representation of a DOM element ? It would be nice if
such algorithm respects the indentation and the presentation (e.g.
underlined or colored text becomes uppercase).
Maybe a kind of editor or filter called from xpcom/C++ and able to
transform HTML elements...