Add open source Piper TTS

[FockeOeuf]

Would you think you'd be able to add the capcity to connect to Piper TTS servers, in addition to cloud services you already implemented?
Those are well known services, which work quite well in local environment, in addition to be free and quite optimized for this use.

[SeriousOldMan]

Do they implement the standard OpenAI protocol for TTS? This is already supported. I don't know Piper to be honest.

[FockeOeuf]

This is quite unlikely. Piper TTS has Python and C/C++ functions to call the server and reverts a bitstream audio feed.
As it is open source, there are extensive documentation and examples about how the interface is built and must be called.

In home assistant environment, Whisper and Piper work with the same Wyoming pipeline. As I can see that you already implemented Whisper, I thought it would be pretty similar to implement Piper TTS. Maybe using Wyoming?

[SeriousOldMan]

I already had a short look. Will be pretty easy to create an integration. Most likely for next Friday.

[SeriousOldMan]

Currently implementing it. Can you do me a favor and run "http://localhost:5000/voices" in your browser and send me what is returned? localhost:5000 is of course the URL of the Piper web server.

[FockeOeuf]

The result is

{"detail":"Not Found"}

[SeriousOldMan]

Hmmm, this is what the docs say:

[FockeOeuf]

Let me check that this afternoon

[SeriousOldMan]

Implementation is done according to the documentation. If you want to give it a try, hold down the Ctrl key and click on Download in "Simulator Startup". Release the key, if the download starts. This will install the (stable) dev version of the upcoming 7.2.2.

If you want to go back, click on Donwload without Ctrl key,

For further discussion and bug hunting it would be best, if you can join our Discord.