There is a new open source Text to Speech generator in town called Kokomo-82M.

As far as I can determine, it’s being developed by one person, Hexgrad, based on earlier models.

Apparently, this is something you can install and run locally on your own computer.

You can try it out online here. You can also compare various open source models at the TTS Arena.

My take: note that this does not clone voices or emote (at all.) Perhaps in the next version?