I understand that we cannot use an Amazon Echo via Alexa Media Player as a media_player for piper to output to as it needs to use a notification service instead of a tts, making it impossible for us to use one of pipers other voices instead of the standard Alexa one.
However it is possible to play an MP3 via the notify.alexa_media service (as an example below)
service: notify.alexa_media
data:
message: "<audio src='https://automatiseer.eu/audio/kleine-wasjes-grote-wasjes.mp3' />"
target:
- media_player.my_echo_dot
data:
type: tts
It is also possible to create an Mp3 file - at least (at the first time that sentence is used) via Cache of the Piper-created TTS which sits by default in config/tts
Would it be possible, and a large step to be able to force that mp3 file to be created each and every time, then uploaded to a public IP that Amazon could see and then play that mp3 file using the above service?
I just thought that would be a good workaround to allow us to use Piper on our Echo devices, although I appreciate it would probably be slower due to the conversion time, and upload then playback.