Skip to main content
The Audio Streams API lets you fork the audio of a live call to a WebSocket server in near real time. You can start and stop streams dynamically during the call, retrieve the list of active streams, and optionally send audio back into the call using bidirectional streaming.

​
Key Capabilities

  • Real-time Audio - Fork inbound, outbound, or both tracks of a live call to your WebSocket in near real time.
  • Bidirectional Mode - Enable bidirectional=true to send audio from your server back into the call.
  • Direction-specific formats - Receive inbound L16 at 8 or 16 kHz, or μ-law at 8 kHz. In bidirectional mode, send outbound L16 playback at 8, 16, or 24 kHz, or μ-law at 8 kHz.
  • Status Callbacks - Receive stream lifecycle events (connected, stopped, timeout, failed) on your callback URL.
Important: Your WebSocket server must accept connections over wss://. Audio packets are delivered as JSON frames containing base64-encoded payloads. See the Stream XML element to start streams from XML instead of the API.

​
API Endpoint

Base URL
Stream Base URI
Authentication required:
  • X-Auth-ID - Your account ID (e.g., {auth_id})
  • X-Auth-Token - Your account Auth Token
  • Content-Type: application/json

​
Available Operations

Note: Audio streams are billed per minute of audio forked. Stopping a stream immediately halts billing for that stream.