Streaming audio output

[fivestones]

Is your feature request related to a problem? Please describe. I'm always frustrated when I try to convert a long text to speech and I have to either use interactive mode (which, in order to play on any system other than the mimic3 server computer, leaves me with a bunch of files) or I have to just wait until the entire thing is converted (like when using curl to post the text to the api and listening to the resulting download.

Describe the solution you'd like When using curl to download audio from the api, this audio should start immediately instead of only starting when the entire text is completed.

Describe alternatives you've considered I tried using the mimic3 command directly, but it seems to do the same--it works until it has finished the whole text, and then outputs the entire audio. If you use --interactive mode it outputs audio after every sentence, which might be fine if I were using the server computer mimic3 is running on to actually listen to the audio, but I would like to be able to listen to the audio on a different machine (e.g., my phone), in close to real time.

Additional context If the speech audio immediately started coming from the api when sending it a post request with the text to convert to speech, curl could be used to stream the audio and pipe it to some other app to play it in real time. This would be much easier than trying to deal with all the files created by --interactive mode, and as far as I can tell interactive mode isn't available anyway with the api (although since the api documentation page is broken it's hard to know for sure).

Thanks so much! The voices sound amazing.