#TTS Providers
1 messages · Page 1 of 1 (latest)
Looking at this PR https://github.com/home-assistant/core/pull/163114
And I didn't quite know how to handle the old TTS Provider. Looked around and found this PR by @tribal pewter https://github.com/home-assistant/core/pull/108293
He writes: "For now, we'll keep the legacy cloud tts provider and set it up to allow users that use the corresponding service to still use that. We haven't decided if and when we want to deprecate that."
Was wondering if that still is the case?
The legacy tts actions are being deprecated in favor of the TTS entities. But that is currently happening on a per integration level. New integrations should only implement TTS entities.
Makes sense. This is not a new integration though and still has a legacy TTS provider class. Should we issue a repair warning - nudging the user to use the TTS entity instead, or should we just support both - as is the case for the cloud integration?
which integration?
The PR I'm reviewing adds a config flow to the "Pico TTS" integration. The other integration I was mentioning is "Home Assistant Cloud".
ah, ok, I also reviewed that PR
don't put too much effort in the tests, there is a second PR that will replace the PicoTTS executable with a python library
At the time we decided to keep the Cloud TTS provider to avoid a breaking change in automations.
I'd say that most integrations can deprecate the old provider with a repair issue.
I see, so raising a repair issue in async_get_engine is fine?