Thanks. There is a lot of trickery going on to keep the token count in check, mainly splitting strings. Feel free to check out the code at the bbs.
