Separate service to generate TTS and play it

[julianrinaldi]

It would be nice if there was a way to call a service to generate a TTS message before actually playing it. My specific use case is kind of ridiculous, but it would be nice to have. I have buttons for different timers in my gym, like 30 seconds, 60 seconds, 90 seconds, etc, that I press to have a rest period before different weight lifting exercises. Then at the end of the timer, I generate a TTS message through Eleven Labs of a voice of Rocky Balboa that generates different inspirational messages in his linguistic style using OpenAI. But both OpenAI and Eleven Labs add a combined delay of between 10 and 20 seconds. I have the timers set to do the TTS 15 seconds before the timer goes off. But it would be nice if it could be generated beforehand and then played at the exact moment the timer goes off. As far as I know this isn't currently possible.

[dbullendsl]

you can use the Chime TTS: Say URL service to do that, it'll create the chime message and give you a URL for the output.

[nimroddolev]

In the next version of Chime TTS, the chime_tts.say_url service will include a media_content_id key value in its response which you will be able to use to call media_player.play_media at a later time, as you described.

For the time being @dbullendsl is correct in that you can use the url key value in the chime_tts.say_url's response so that the notification is ready to be played at a later time.

[nimroddolev]

@julianrinaldi, with the release of v1.0.0 you can now prepare a TTS ahead of time, per your request.

To start using it, make sure to first configure your chime_tts.say_url folder path to be inside your /media folder. You can then call the chime_tts.say_url service, which will return the media_content_id of the generated audio for you to use in your automations.

[julianrinaldi]

Awesome, thanks. I'll give it a try.