Tts.cloud_say with variables reads out the variable name

Hi all,

I’m setting up a simple TTS notification using the Nabu Casa voice system and I’m really impressed with the voice feedback compared to the native Google TTS I’ve used previously. I had a couple of these working with the google tts using NodeRed but stripped them out as I moved all my automations back to yaml and the front end (where possible). Something I don’t think I ever tested though, was feeding a variable into the text. It looks like it should be simple enough, and my twitter notifications already do it, but I just can’t seem to find the right syntax so the tts_cloud_say call reads out the variable name instead of the contents.

In this example, I’m trying to pull in the current temperature.

The dev tools template editor confirms the variable is pulling in the correct value but in the message it’s just read out as text. I’ve tried the following options but I get the same results

"Good afternoon. The current temperature is {{ states('sensor.audenshaw_temperature') }} degrees."
Good afternoon. The current temperature is "{{ states('sensor.audenshaw_temperature') }}" degrees.
Good afternoon. The current temperature is '{{ states('sensor.audenshaw_temperature') }}' degrees.
Good afternoon. The current temperature is "${{ states('sensor.audenshaw_temperature') }}" degrees.
Good afternoon. The current temperature is '${{ states('sensor.audenshaw_temperature') }}' degrees.

I’ve been hunting through a few examples and GitHub repo’s with similar tts notifications and I don’t see anything different.

What am I missing folks?

Any news? I have the same problem.