looking at HTML::TokeParser's pod, looks like just a simple call to the get_trimmed_text() method will get what you want.. be sure to look at the two examples in the docs there -- i think they're both applicable...