Wyoming Protocol no devices?

[fluid moth]

I am trying to configure a esp32 device that is working as a speak assistant. For this to work i downloaded the Openwakeword, Whisper and Piper add-on. All logs are not stating an error. The only thing is all the add on entities are either status unknown or not available.
I can't figure out how to fix it.

[elder gazelle]

what are you running haos on?
is your home assistant up to date?
is there anything in the addon logs that might hint at why they are not running?
what esp32 device are you running? custom?

[fluid moth]

what are you running haos on?
A VM in Proxmox, 8GB RAM. AMD Ryzen 7 5825U
is your home assistant up to date?Installation method Home Assistant OS
yes Home Assistant is up to date:
Core 2026.1.0b4
Supervisor 2026.01.0
Operating System 17.0.rc
is there anything in the addon logs that might hint at why they are not running?
Only in whisperer: [23:51:38] WARNING: Your CPU does not support Advanced Vector Extensions (AVX). Whisper will run slower than normal.
Piper: 2025-12-30 19:43:16.866598607 [W:onnxruntime:Default, device_discovery.cc:164 DiscoverDevicesForPlatform] GPU device discovery failed: device_discovery.cc:89 ReadFileContents Failed to open file: "/sys/class/drm/card0/device/vendor"
what esp32 device are you running? custom?
esp32-s3-devkitc-1

[elder gazelle]

I think your CPU type setting of your VM is not set correctly in proxmox
it should be set to "host"

[fluid moth]

unfortunately this didn't solve my problem 🙁

[elder gazelle]

you done a "hard" vm restart since changing?
are the addons still failing? has anything changed in their logs?

[fluid moth]

I pressed shutdown and then started the VM again.

Logs:

Log Piper: 
INFO:__main__:Ready
[11:42:24] INFO: Successfully send discovery information to Home Assistant.
s6-rc: info: service discovery successfully started
s6-rc: info: service legacy-services: starting
s6-rc: info: service legacy-services successfully started
~~
Log Whisper:


INFO:__main__:Ready
[11:42:28] INFO: Successfully send discovery information to Home Assistant.
s6-rc: info: service discovery successfully started
s6-rc: info: service legacy-services: starting
s6-rc: info: service legacy-services successfully started


Log OpenWakeword


INFO:root:Ready
[11:42:24] INFO: Successfully sent discovery information to Home Assistant.
s6-rc: info: service discovery successfully started
s6-rc: info: service legacy-services: starting
s6-rc: info: service legacy-services successfully started

[elder gazelle]

can the wyoming integration see the addons?

[fluid moth]

Yes

But they aren't grouped, don't know if that is a problem?

[elder gazelle]

in the voice assistant settings menu you need to create a voice pipeline. then you assign it to your device.
NB: if your using OWW then you will need to add a streaming wake word in the overflow menu. but if your using an ESP32-S3 then you are better off running microwakeword on the device instead.

how is your device set up?
given you said esp32-s3-devkitc-1 to my earlier question I assume its a custom device?
what does the circuit look like and what does your ESPHome config look like?

[fluid moth]

I am using the ESpressif ESP32-WROOM -S3- DevkitC1, PSRAM16MB, N16R8 Cp2102

ESP Config:

  name: esp-jarvis-s3
  friendly_name: ESP Jarvis S3

esp32:
  board: esp32-s3-devkitc-1
  framework:
    type: esp-idf

logger:
  level: DEBUG

api:
  encryption:
    key: "6JpW8kL2mN4pQ6rS8tV2wX4zA6bC8dEfGhIjKlMnOpQ="

ota:
  - platform: esphome

wifi:
  ssid: "SSID"
  password: "Password."

i2s_audio:
  - id: i2s_shared
    i2s_lrclk_pin: GPIO12
    i2s_bclk_pin: GPIO11

microphone:
  - platform: i2s_audio
    id: board_mic
    i2s_audio_id: i2s_shared
    i2s_din_pin: GPIO10
    adc_type: external
    pdm: false

speaker:
  - platform: i2s_audio
    id: board_speaker
    i2s_audio_id: i2s_shared
    i2s_dout_pin: GPIO9
    dac_type: external

voice_assistant:
  id: va
  microphone: board_mic
  speaker: board_speaker
  use_wake_word: true
  noise_suppression_level: 2
  auto_gain: 31dBFS
  volume_multiplier: 2.0

# Voeg een switch toe om wake word handmatig te bedienen
switch:
  - platform: voice_assistant
    name: "Use Wake Word"
    id: use_wake_word

light:
  - platform: status_led
    id: status_light
    pin: GPIO48````

I attached the pipeline config.

Sry if I am providing wrong info

[elder gazelle]

are you copying a project from somewhere?
why use a shared i2s bus instead of a separate i2s bus for mic/speaker?
why are you wanting to use OWW instead of on device MWW?

if you have reasons for doing it that way its cool, but it seems like you may be basing some stuff on outdated examples.

[fluid moth]

First esp32 project and used gemini for the yaml code

[elder gazelle]

how about the wiring though where you decided to share the i2s bus?

I am guessing this is based on an older project and also ai's are going to have a lot of older stuff in the source data.

voice has been evolving very quickly and guides get out of date very quickly.

[fluid moth]

I just followed the installation scheme of gemini. The shared esp picture is the one I have.

[elder gazelle]

but what is that? what did it tell you to do when building it? i am more curious then anything else.
if you don't explain the exact hardware setup you have then people are going to struggle with helping you with firmware configuration.

what audio stuff is wired to it and how?

[fluid moth]

I am using a INMP441 for the listening part of the ESP and MAX98357A for the audio amplifier

[elder gazelle]

do you have a circuit diagram?

[fluid moth]

No I haven't. I think I underestimated this project a bit 😅. Think I will read a bit more about how to fix it. Thank you for your help

[elder gazelle]

you might be better off just getting a Voice-PE but learning is always good. there is also the ESPHome discord server which may be able to help a little more as you learn than here - https://discord.gg/JsyZmvZm

[fluid moth]

Thanks for the info!