Allow streaming audio responses

[BeeShall]

Is your feature request related to a problem? Please describe.
Whenever I want to use text to speech as a chatbot response, I have to wait for the response from the LLM and then use a TTS service to synthesize the results which results in a huge latency and poor UX experience.

Describe the solution you'd like
It would be really cool is cl.Audio/cl.Message allowed streaming of audio tokens so that we can stream TTS response as we get it, instead of waiting to synthesize completely.

Describe alternatives you've considered
None at the moment.

[chibexme]

I'm looking forward for this functionality as well.