2 rpi 3+
2 psEye cameras for the mic arrays (bought when i was working with snips)
2 additional USB mic+headphone adapters.
I already have an MQTT broker on my network for alot of my home automation stuff. I am guessing though that rhasspy would be like snips and could potentially send alot to the mqtt broker (audio frames), but I saw a recommendation in another post that maybe the listeners have an onboard mqtt broker for the audio frames… Not sure if I could still use my existing fault tolerant mqtt installation as the master.
One other question - I definitely want to work with the master/satellite setup. My plan is to run this on the pi’s i already have around the house for various things. The master however, I am considering putting on something a little more powerful. I have an esxi setup here at home and could have my master node running on a VM with more power. Has anyone done this with running the master on something that has no audio hardware but just using it for processing?
Thanks! Excited to get moving with this after the let down from sonos/snips
Yes, I have Home Assistant running on a VM on proxmox (NUC). This has the Mosquitto addon installed and is used as master. HA is also running the Rhasppy Addon, connected to the Mosquitto.
Yup, the Rhasspy speech-to-text instance (‘server’) runs on the NUC i3 Home Assistant system that’s tucked away upstairs with other computers, no microphone, no speaker. Two RPi3 satellite (‘clients’) are placed downstairs in the living room and upstairs in the bedroom. Not using MQTT at all. Just Rhasspy doing it’s thing in the background.