What's winders? Windows? If you're asking a question, don't throw in stupidisms. The harder it is for me to parse your question, the less likely I am to answer it.

Assuming you want to parse each text file for things that look like sitenames ... this is an easy (if time-consuming) task in Perl. Try something like:

while (<>)
{
    open IN, $_ || die $!;
    while (<IN>)
    {
        if (m!https?://([^/"'\s]+)!)
        {
            print "'$1'\n";
        }
    }
    close IN;
}
[download]

This code is untested. Please try and figure out what it's doing before you tell others it works. It also isn't complete and is easily fooled. It's an 80/20 solution, useful only if 80/20 solutions are acceptable. Much better would be to employ something like URI or HTML::Parser.