Helo, Bastian (author of webcleaner), bob (filterproxy) and me (aifilter) are currently discussing these things via email. I invite evryone to have a look at bloodgate.com/aifilter (and especially at the documents inside doc/performance of the downloadade source) to see how I solved the problem with HTML::Parser. The parser is fast, but if you need it to re-parse the document 10 times it starts to get real slow. Of course, bob is right, regexps can be faster than the parser. Anyway, suggestions, opinions, comments etc to this site, or via email. Thanx in advance, Tels (bloodgate.com)