Freeswitch - Google Speech to text

We would like to create a voicebot for our company’s internal needs.

Our idea is that an incoming call will be routed to FreeSWITCH, which will then real-time forward the call to Google Speech-to-Text API. We will further process the transcribed text and then return the response (either from Google or our own program) back as a stream.

We attempted to install and configure it ourselves, but we were unable to successfully forward the stream to a Node server (we are unsure if this is even necessary).

Therefore, we are asking for your assistance and a brief summary of what needs to be done.

I would like to get a basic understanding of how to best install and configure the entire system.

1) FreeSWITCH Community or SignalWire variant?
2) What additional modules, not included in the default installation, are required?
How is stream forwarding handled? Through dialplan? What module or method is best to use?
3) Could you provide an example of a dialplan?
4) Is it necessary to have a separate Node.js or Python server to forward the stream to Google API, or can FreeSWITCH directly send it to Google API? If so, how?

Thank you very much.

Dušan.