wvZopeFilter tries to remove many purely "appearance-related" HTML tags from HTML files written by wvWare or unrtf and to keep only structural information from the documents.
Since at least two good Zope Products to handle Word doc files (NuxDocument and MSWordDocument) already exist, I decided not to implement my own complete Zope product, but "only" a little module, which can be used to post-process the HTML data return for example by NuxDocument.getHtml(). Moreoveer, it can be used a as stand alone program.