Holy Frick! 11labs quality and fast speed TTS finally all local!
Wow this post blew up! Just wanted to point out:
The repo below isn't mine, I have an audio sample on my fork, in stall from kanttouchthis, their repo is compatible with windows now.
This is the extension I'm referencing:
https://github.com/kanttouchthis/text-generation-webui-xtts
I got it working on a windows installation, here is an issues for more information:
https://github.com/kanttouchthis/text-generation-webui-xtts/issues/3
Two things to note* obsolete now:
-
reference the code change to fix the auto play issue if you are having one.
-
and very importantly, I think this is a windows only thing, change the install folder (in the extensions directory) from
text-generation-webui-xtts
to
text_generation_webui_xtts
It totally works as advertised, it's fast, you can train any voice you want almost instantly with minimum effort.
Abide by and read the license agreement for the model.
**Edit I guess I missed the part where the creator mentions how to install TTS, do as they say for the installation.
https://github.com/RandomInternetPreson/text_generation_webui_xtt_Alts/tree/main#example
Example of output, took about 3 seconds to render after the ai had finished the text.