Is it possible to control the text more?

[maniac-0s]

The speech currently is just running the text down, depending on the voice model more or less fast. I wish one could control the text more. "..." for instance seems to be ignored most of the time just read like a single dot.

Is there any way to implement controls into this, maybe via tags in a prompt?

[64jcl]

I have had some success replacing the ... with ; (a semicolon) - but sometimes it just ignores that even. There really should be some way to make it actually pause on those.