gilded cedar
Hey,
Can anyone tell me which software/models are used to power HA Cloud's speech-to-text and text-to-speech features?
Following the presentation from the HA Voice PE release, I've seen the comparison list between the languages supported locally compared to those with HA Cloud and was wondering why the difference is so big. I'm sure the cloud runs on stronger hardware and, as such, can use larger models that provide better accuracy but some of the languages there aren't supported at all with Piper so I'm guessing something else is used.
Thanks in advance!