Skip to main content
Question

Enabling bi-directional realtime audio streaming though incoming call

  • January 20, 2026
  • 4 replies
  • 156 views

An incoming call to a RingCentral phone number should be routed to our Next.js application (running server-side with WebSocket support), where the real-time audio media stream should be received. Our application then processes the audio, which should be sent back into the same RingCentral call.

In effect, the call audio should flow bi-directionally through our application, acting as a real-time media processor.

How can we achieve this requirement through API?

4 replies

PhongVu
Community Manager
Forum|alt.badge.img
  • Community Manager
  • January 20, 2026

Check out this blog article to learn how to achieve that.


  • Author
  • New Participant
  • January 20, 2026

@PhongVu 

Thank you for the reply.


Instead of using SIP connection, can we achieve this requirement through API like
Voice API or any other bi-directional streaming APIs or SDK.
Also, by using APIs provided by RingCentral Contact Center (RTCC) or RingCX license.


PhongVu
Community Manager
Forum|alt.badge.img
  • Community Manager
  • January 20, 2026

Unfortunately, we don’t have such a voice API to stream bi-direction audio. We have the supervision API for inbound audio stream for listening. But it requires a device ID of a SIP device or an WebRTC device.
 

The capabilities mentioned above are only for RingEX. RingCX API is very limited and I don’t think that it is supported. I am not an expert in RingCX platform, you can browse this dev guide to see if it’s possible at all https://developers.ringcentral.com/engage/voice/guide/bots. You can post your questions relating to RingCX to the “Contact Center & RingCX” group


  • Author
  • New Participant
  • January 26, 2026

@PhongVu 

Thank you for your reply.

Since the Softphone requires connection to be associated with an individual user’s phone number, incoming calls must first land on a specific user's phone number. Our AI receptionist then answers the incoming call via the Softphone implementation, processes the request, and routes the call to the appropriate call queue or extension.

Currently, this implementation with Softphone is limited to use the user's phone number only, which only supports five concurrent calls. This is insufficient for our primary reception line.

Could you please clarify if there is a way to leverage the Softphone SDK to support more than five concurrent calls? Alternatively, is there a recommended approach to achieve this high-volume requirement for our AI integration?