Logs on all 4 are great and displaying they are ready, like this:
INFO:__main__:Downloading FasterWhisperModel.TINY_INT8 to /data
INFO:__main__:Ready
The one running on the regular ports 10200 and 10300, I am able to connect through the Wyoming integration with no issues, and they are working. But the one I mapped to nonconflicting ports 10201 and 10301, I am unable to connect with the Wyoming integration.
The first port is the port on the host, the second port is the port in the container. This container always opens port 10200, so you need to change your mapping to:
Could be the assist integration. All my windows and doors have Xiaomi sensors on them which show as binary_sensor which currently is not supported by assist
Very cool stuff. Can we/how can we, train Piper with our own voices? That way we can have our own voice respond to the voice prompts. For example, you could have mom or dad respond to certain prompts as defined in the pipelines.
They should really update this blog post to specify that Whisper and Piper on a Raspberry Pi 4 is only supported, at least when running Home Assistant OS, on a 64-bit version of the OS (which was not available when I initially installed HA). If you are still running 32-bit like I was, the add-on installation page will tell you these add-ons are not compatible with your operating system.
Maybe it is asked before but, how to get the show on the road?
I see two add-ons, Whisper and Piper
I see one integration in add integrations, Wyoming protocol
I like to use it with rhasspy 3 which should be able to work with Wyoming protocol.
Do I need to install Whisper, Piper and the integration of Wyoming Protocol or does Whisper include that automatically or Piper does that. WHat is need to get it running.
Or is everything already neatly installed in the latest update of HA?
O and one last thing I see all these intents that you need to place in the configuration.yaml, would it be possible to place all voice related things in one other included yaml file, for it clutters the configuration file so badly, specially if you like a lot of Intents?
Can I download 2 or more voice assistants locally? I have seen examples of having a local and the HA cloud but haven’t seen anything mentioning having 2 local ones. There is an example above when using docker however since I am installing the add-ons in HA how would this be done? Do you need to load multiple instances of the Piper & Whisper add-ons or is it just a multiple instance of the Piper add-on? I want to have 2 different languages available locally.
I just realised I used the example setup for piper and had a dutch voice loaded. I have changed to en-gb-southern_english_female-low but still piper is greyed out…
To answer your return question, yes I did, was following this post a year ago when WHisper and Piper were not the route on youtube to follow then it was Rhasppy, so I bought a for me expensive conference speaker and wanted to start, then came Whisper and Piper, the wizards of HA decided that would be the route.
And yes Rhasspy 3 will work with the Wyoming protocol, stil that does not answer my questions, nor your return question did, for how will they work together, do I still need whisper/Piper or can I go with Rhasspy and install the wyoming protocol via add-ons?
So I think my question has not been answered still not clear what way to commence!
Plus the other question was is it possible to put all the Voice stuff in one Yaml file, or the intents only for when there are a lot of integrations and entities you like to control your COnfiguration. yaml will get heavily cluttered, I like to decentralise all things from Configuration.yaml, way more structured, way easier to edit and find if you want to add or edit, and if one edit goes wrong only that one will be bricked not your whole HA.