Piper add-on custom voices cannot be selected in Settings > Addons > Piper > Configuration

[domesticatedviking]

Describe the issue you are experiencing

I have some custom piper models that I built with TextyMcSpeechy which I want to use with my automations in home assistant. These are stored in /share/piper and they appear correctly in Settings > Voice Assistants when I create Assistant entities. They also produce speech properly when tested under that system.

I want to be able to use these voices in my own automations, but there is no mechanism for adding them to the dropdown in Addons > Piper > Configuration > Voice, which determines the system-wide voice.

I presume this is because this list is hardcoded somewhere like config.yaml which is located inside the Piper add-on's docker container, which cannot be edited from the Home Assistant OS webui.

Edit: the Wyoming integration may also be a possible culprit here as it seems to be downloading its own list of voices in voices.json from huggingface.

DEBUG:__main__:Namespace(piper='/usr/share/piper/piper', voice='en_US-lessac-medium', uri='tcp://0.0.0.0:10200', data_dir=['/data', '/share/piper'], download_dir='/data', speaker='0', noise_scale=0.667, length_scale=1.0, noise_w=0.333, auto_punctuation='.?!', samples_per_chunk=1024, max_piper_procs=1, update_voices=True, debug=True)
DEBUG:wyoming_piper.download:Downloading https://huggingface.co/rhasspy/piper-voices/resolve/v1.0.0/voices.json to /data/voices.json
DEBUG:wyoming_piper.download:Loading /data/voices.json
DEBUG:wyoming_piper.process:Starting process for: en_US-lessac-medium (1/1)
DEBUG:wyoming_piper.download:Checking /data/en_US-lessac-medium.onnx
DEBUG:wyoming_piper.download:Checking /data/en_US-lessac-medium.onnx.json
DEBUG:wyoming_piper.process:Starting piper process: /usr/share/piper/piper args=['--model', '/data/en_US-lessac-medium.onnx', '--config', '/data/en_US-lessac-medium.onnx.json', '--output_dir', '/tmp/tmpmjmisltq', '--json-input', '--noise-scale', '0.667', '--length-scale', '1.0', '--noise-w', '0.333']
INFO:__main__:Ready

It would be useful if :

1. the add-on would be made aware of models located in /share/piper on startup.
2. the list of voice models available to the piper add-on could be governed by a config file located outside the docker container which could be edited by the user. (This list is also very long- it would be appreciated if it could be customized.)

What type of installation are you running?

Home Assistant OS

Which operating system are you running on?

Home Assistant Operating System

Which add-on are you reporting an issue with?

Piper

What is the version of the add-on?

1.5.2

Steps to reproduce the issue

1. Add custom voice model with proper file name format (eg en_US-myvoice-medium.onnx and en_US-myvoice-medium.onnx.json) to /share/piper from within the Home Assistant OS webui.
2. Verify that the voice model appears and generates speech in Settings > Voice assistants by creating an assistant entity and using the voice test feature.
3. Restart the Piper add-on.
4. Go to Settings > Add-ons > Piper > Configuration > Voice dropdown menu
5. en_US-myvoice-medium does not appear in the dropdown menu, therefore cannot be selected for use as the system Piper voice.

System Health information

System Information

version	core-2025.1.4
installation_type	Home Assistant OS
dev	false
hassio	true
docker	true
user	root
virtualenv	false
python_version	3.13.1
os_name	Linux
os_version	6.6.62-haos-raspi
arch	aarch64
timezone	America/Vancouver
config_dir	/config

Home Assistant Community Store

GitHub API	ok
GitHub Content	ok
GitHub Web	ok
HACS Data	ok
GitHub API Calls Remaining	5000
Installed Version	2.0.5
Stage	running
Available Repositories	1575
Downloaded Repositories	4

AccuWeather

can_reach_server	ok
remaining_requests	15

Home Assistant Cloud

logged_in	false
can_reach_cert_server	ok
can_reach_cloud_auth	ok
can_reach_cloud	ok

Home Assistant Supervisor

host_os	Home Assistant OS 14.2
update_channel	stable
supervisor_version	supervisor-2024.12.3
agent_version	1.6.0
docker_version	27.2.0
disk_total	56.3 GB
disk_used	7.3 GB
healthy	true
supported	true
host_connectivity	true
supervisor_connectivity	true
ntp_synchronized	true
virtualization
board	rpi4-64
supervisor_api	ok
version_api	ok
installed_addons	Duck DNS (1.18.0), File editor (5.8.0), Terminal & SSH (9.16.0), Mosquitto broker (6.5.0), NGINX Home Assistant SSL proxy (3.11.1), motionEye (0.21.0), Cloudflared (5.2.7), Piper (1.5.2), FTP (5.1.1)

Dashboards

dashboards	1
resources	3
views	10
mode	storage

Recorder

oldest_recorder_run	August 5, 2024 at 3:06 AM
current_recorder_run	February 1, 2025 at 9:25 PM
estimated_db_size	114.25 MiB
database_engine	sqlite
database_version	3.47.1

Anything in the Supervisor logs that might be useful for us?

Anything in the add-on logs that might be useful for us?

Additional information

No response

[domesticatedviking]

This thread is probably useful: https://community.home-assistant.io/t/how-to-install-custom-piper-voice-model-on-haos-installation/617267/5

[domesticatedviking]

While I would still like this issue to be addressed, after some further exploring I found I was able to make use of my custom voice in my own automations by doing the following:

1. I copied en_US-mycustomvoice-medium.onnx and en_US-mycustomvoice-medium.onnx.json to /share/piper using the FTP add-on and the Filezilla FTP client on my PC.
2. I restarted the Piper add-on and reloaded the Piper entity in the Wyoming integration.
3. I created a new home assistant script using the yaml below (substitute your own media player entity id and voice name for media_player.myspeakername and en_US-mycustomvoice-medium)

data:
  cache: true
  message: This is a test message
  media_player_entity_id: media_player.myspeakername
  options:
    voice: en_US-mycustomvoice-medium
target:
  entity_id: tts.piper
action: tts.speak

Calling this script successfully renders the message in the custom voice.

[domesticatedviking]

To simplify the use of custom voices and make it possible to call them like any other TTS service,
I created a service script to hold the configuration for each voice which could receive messages as a template variable.

Service script: say_as_bob (note that you need to use your own media_player_entity_id and voice )

action: tts.speak
metadata: {}
data:
  cache: true
  message: "{{ message }}"
  media_player_entity_id: media_player.myspeakername
  options:
    voice: en_US-bob-medium
target:
  entity_id: tts.piper

To make Bob say whatever you like, make scripts like introduce_bob below (this yaml also works from developer tools > actions)

action: script.say_as_bob
data_template:
  message: "Pleased to meet you, I'm bob"

[domesticatedviking]

For anyone who is here because they are struggling with this issue, I just wrote a fairly comprehensive guide to using custom piper voices in Home Assistant OS. It can be found here.