possible to stream live audio to a websocket using uniMRCP

[Shara Karasic]

Hi,

We already use uniMRCP with Google Speech for speech recognition. We just got a customer request to stream live audio from a call to a websocket. Seems like it would be possible to do this with uniMRCP - can we do this using uniMRCP? Is there documentation on how to do this?

Thanks for any info!

Best,

Shara

P.S. I am a PM, not a developer.

[Jones Luk]

A simple answer is yes.

uniMRCP doesn't have good document.

But you can look at $unimrcp-source-root/plugins/mrcp-recorder/src/mrcp_recorder_engine.c, this guy save audio frames into a file when receive. You should be able to customize it so that frames are sent to websocket instead.