This should be simple with HTML::Parser.

See the Extract all plain text from an HTML file example to start with, then read the main docs to "fine tune" things.

HTH.