Linux Voice Assistant project from OHF voice lead dev is successor to the previous "Wyoming Satellite" project

[Hedda]

FYI, the Open Home Foundation voice lead developer recently launched a new Linux Voice Assistant project that is bastically an experimental Linux voice assistant for Home Assistant that uses a port of the ESPHome protocol, with goal to support Linux-hardware like ex. Raspberry Pi.

• https://github.com/OHF-Voice/linux-voice-assistant

As I understand it this is kind of a a successor to the Wyoming Satellite project which in practice was put on hold due to shifting focus.

Full explaination here:

• Replace Wyoming satellite with Linux Voice Assistant OHF-Voice/backlog-issues#45

Also read:

• https://www.reddit.com/r/homeassistant/comments/1na1chk/new_wyoming_satellite_successor/

Quote:

Author here, glad to see so much interest :)

To give some context: the Wyoming protocol was developed for voice services like speech-to-text, text-to-speech, etc. I added satellite messages early on so we could use Raspberry Pi's alongside the ESPHome devices. Over time, the Wyoming Satellite feature set has fallen behind ESPHome and it's felt like I'm basically recreating the ESPHome messages in Wyoming.

So I had the idea: why not just implement a tiny subset of the ESPHome protocol on Linux and reuse what's already there? This turned out to be way easier than I expected! It also has the benefit that I can expose the media player properly now, and it will be possible to expose LED rings and buttons too in the future without any changes to Home Assistant. The Raspberry Pi satellite just looks like an ESPHome device with the proper entities already in place.

At the moment, this project only supports microWakeWord but I'd like to add openWakeWord support soon (PRs welcome). I'm also using the mpv media player directly now for audio output so I can mix audio streams properly. Things like timers, announcements, start conversation, audio ducking, etc. all just work out of the box. It should also be possible to use PipeWire/PulseAudio plugins for acoustic echo cancellation (AEC) so the wake word can still be heard over music.

The next step will be to add a proper tutorial and start collecting feedback from beta

Its counterpart is a Python Client for ESPHome native API used by Home Assistant:

• https://github.com/esphome/aioesphomeapi

PS: So as I understand it, this is meant to replace their previous "Wyoming Protocol" which is a peer-to-peer protocol for voice assistants:

• https://github.com/OHF-Voice/wyoming

As such it will be used in new remote voice satellites that previously used the Wyoming protocol and was therefor called a "Wyoming Satellite":

• https://github.com/rhasspy/wyoming-satellite

[Hedda]

@j1nx @BohdanBuinich @goldyfruit @JarbasAl just trying to brainstrom but would like to believe that the OHF might be interesting in coolaborating on a project to write an open-source linux kernel driver for the XMOS XU316 chip which could be based on your VocalFusionDriver code if you add a license?

• [IDEA] Write a fully free and open-source driver for the XMOS XU316 / XVF3610, XVF3500 and/or XMOS XVF3800 if possible? OHF-Voice/linux-voice-assistant#17

• [REQUEST] Add a "free and open-source" license to this VocalFusionDriver repository, project and code? VocalFusionDriver#13

[Hedda]

Very much related to the new article on a using Home Assistant as a backend stack that JarbasAl written for the the Open Voice OS blog here:

• https://blog.openvoiceos.org/posts/2025-09-17-ovos_ha_dream_team

[JarbasAl]

thanks for sharing, always good to know what others are up to and what sort of issues/benefits different approaches bring to the table

my solution for satellites/hub-and-spoke is hivemind https://jarbashivemind.github.io/HiveMind-community-docs/ , in the post i didn't really explain what hivemind is has that was not the focus

in the end neither wyoming nor it's replacement are very useful to OVOS, since OVOS itself is an assistant but in the HA world a "voice assistant" is just a smart speaker device

i.e. in OVOS world the "assistant" is the brain, in the HA world the "assistant" is the satellite hardware

at some point I might look into implementing the hivemind protocol into my Voice Pe :)