there are many ways to do this in batch.
(1) regular expression. Not the best way; very prone to error if html is malformed.
(2) Parsing the html with a specialized parser, e.g., perl's HTML:Tree or Ruby's REXML. More accurate, but still a pain for a simple plain text dump.
(3) (recommended) use a text-only browser (e.g., lynx, links). Pipe the whole set of files into say lynx. It has an option to dump text files (I think it was -d). This is the easiest and most robust.
You need to download lynx from: lynx.browser.org/