Offline Home Automation and Voice Assistant

Hello Mycroft friends!

I apologize if the following question that I am asking is very basic. I am fairly new to self-hosting, engineering, and all things tech, and was wondering if I could get some assistance!

TLDR Looking to set up my own smart speaker system for music. Specifically using good quality speakers that function independently from the smart service provider itself.

The goal would b e to have essentially three devices running mycroft front end with a central locally hosted mycroft core server running on more powerful hardware (3700x and 3070, w/ 16gb 3200 c16 ram)

Questions are:

  1. Are multiple clients with a “remote” processing center, something that is supported
  2. Is there a way to easily set this up
  3. Is there a generally easy setup for a completely locally hosted instance of Mycroft?

From my understanding there are 4 main components
a. Wake Command
b. Speech to text
c. Processing / utilizing the skill
d. Text to speech.

A. i know is something that can be done locally
B. Im unsure of
C. Most of the skills I would do this for are calling out to local services, so no worries there
D. Mimic 3 is now locally entirely.

Is there an easy guide / best pathway for getting this set up locally?

1 Like

For Speech-to-text there is a usable english model for Coqui-STT. As you have a decent GPU you probably want to have a look into the STT models for Nivida Nemo as well.

For the centralized Mycroft server look into Hivemind