Skip to main content

Voice and Video Calls

Voice and video calls use LiveKit for WebRTC. The backend exposes api/Call for tokens and call lifecycle; the frontend uses the LiveKit client and SignalR for call signaling (initiate, accept, reject, end).

Overview​

  • Backend: api/Call — get LiveKit token, initiate call, accept, reject, end. Call initiation is also signaled via SignalR so the callee receives an incoming-call event.
  • LiveKit: Room-based; each call has a room name. Direct calls use a room like call_{callerId}_{receiverId}_{ticks}; group calls use groupcall_{groupId}_{ticks}.
  • Push: When the callee's tab or app is in the background, they can receive Web Push or Expo Push for incoming calls (see Web Push and Expo Push).

API Endpoints​

MethodEndpointDescription
POST/api/Call/tokenGet LiveKit token for a room (body: roomName, optional displayName, canPublish, canSubscribe)
POST/api/Call/initiateInitiate a call (body: receiverId, optional callType e.g. "video", isGroupCall); returns roomName, token, url
POST/api/Call/acceptAccept an incoming call
POST/api/Call/rejectReject an incoming call
POST/api/Call/endEnd an active call

Flow​

  1. Caller: Call POST /api/Call/initiate with receiverId (and optionally isGroupCall, callType). Backend creates a room name and returns a LiveKit token and URL.
  2. Signaling: Backend notifies the callee via SignalR (and optionally RabbitMQ for push). Callee sees incoming call UI.
  3. Callee: Accept → POST /api/Call/accept; then frontend gets a token (e.g. via same initiate flow or a dedicated token endpoint) and joins the LiveKit room.
  4. Both: Connect to LiveKit with the token and join the room for audio/video.
  5. End: Either party calls POST /api/Call/end; UI and room are cleaned up.

Configuration​

  • LiveKit: Configure LiveKit:Url, LiveKit:ApiKey, LiveKit:ApiSecret in the backend (e.g. appsettings or environment). Default dev URL is ws://localhost:7880.
  • Frontend: LiveKit client and optional env (e.g. LiveKit URL) for the client SDK.