can this be used to implement real-time transcription?

[Shmoji]

By real-time, i mean the code detects voice audio and then in the next few seconds it returns text of the transcribed audio as the audio is detected.

I see this is what is happening in the demo video, but im curious if all the code in this repo can be used to reverse engineer what you have in the demo there.

The docs make it pretty clear how to use the cli, but not totally sure how to get the functionality seen in the demo to work in python