I’ve made a small addon that is a wyoming protocol proxy for cloud TTS providers. For now, it supports only Google Cloud and OpenAI, but it can be extended to other providers as well.
It basically gives you the same streaming functionality as you get with the latest wyoming-piper but with cloud TTS providers. It also works with Home Assistant Voice Preview Edition devices. With this, you can ask your LLM to tell you a long story, and you will hear the audio almost immediately instead of waiting for the whole pipeline to finish (provided your LLM integration supports streaming).
You can check it out here: GitHub - eslavnov/wyoming-cloud-streamer: Wyoming protocol server for cloud TTS engines. Curious to hear how it works for you if you try it!