Piper TTS Voices

A while back, while playing with tools to record and release the StoryHack podcast, I discovered Piper TTS as a way of generating audio.

Most of the voice models available had what I considered to be restrictive licenses, so I set out to learn how to train my own. Well, like 6000 hours of figuring things out later, I trained a bunch of them. They are all in as open a license as I can manage. That means some are public domain, and some are other relatively open licenses, as dictated by the data sets I found. The datasets I put together were from various Librivox recordings.

Here is the page where they are available. https://BryceBeattie.com/files/tts