> I'm running a dynamic website with a custom search engine 
> and dynamically loaded videos (by script files) in it. Is 
> it possible to make an offline copy of this website using 
> httrack? 

Well, parts of it, sure.

> And would it be possible to: 
> - remove all references to the search engine (it is 
linked 
> on every page in a fixed menu entry) 

Not using httrack - you'l have to post-process pages using 
some basic bash-style scripting

> - transform the movies into real video files? So urls 
like 
> <http://www.site.com/videoscript?video=blah> should 

Err, not if the videos are streamed. In the countrary, just 
add the following scan rule:
+www.site.com/videoscript?video*

You may have to use also something like:
--assume videoscript=application/x-mpg