If the tags don't nest, then it seems that you could get by with just start tags:

<stage right> Enter Juliet
<speech Juliet> Romeo! Romeo! 
Where the heck are you!
<audience> claps.
[download]

The content after each tag is delimited by the next tag or the EOF. You could decompose this format into its respective parts using, e.g.,

while (<>) {
   if (/^<(.+?)>(.*)$/i) {
      $tag = $1;
      $part{$tag} .= $2;
   }
   else {
      $part{$tag} .= $_;
   }
}
[download]

-Mark