Use pyannote for realtime audio streaming with websocket.

[Aliw7979]

Hello, I implemented a websocket server and it receives realtime audio chunks from user in bytes.
Is there any way to send bytes of audio to pyannote for speaker diarization?
Mention that I used diart for this purpose but it cant support multiple user.
Even this link from other similar discussions but couldn't help me either, because it only used voice activity detection, but I need to get something like this for result :
start=0.2s stop=1.5s speaker_0
start=1.8s stop=3.9s speaker_1
start=4.2s stop=5.7s speaker_0
Thanks.

[hbredin]

Nope -- pyannote.audio does not support streaming speaker diarization for now.
diart is the way to go for now (not sure what you mean by "can't support multiple user").

[Aliw7979]

Thank you buddy, the problem with diart is something that I should ask from their rep & maybe it is irrelevant for here
(In a nutshell, main problem that I faced with is their websocket source to receive audio chunks).
By the way, is there any plan to add streaming speaker to pyannote in 2023?