Skip to main content
Astra Dial supports bi-directional audio streaming over WebSocket, enabling real-time audio exchange between your telephony system and external endpoints. This is used to send caller voice data to a processing endpoint (e.g., a voice bot) and receive processed audio back for playback to the caller.

​
Audio specifications

PropertyValue
Encodingaudio/x-mulaw (G.711 PCMU)
Sample rate8000 Hz
Bit rate64 kbps
Bit depth8-bit
FormatBase64-encoded PCM mono audio
Message frequencyEvery 100ms
Audio payloads must be at least 160 bytes or a multiple of 160 bytes (e.g., 320, 800, 4000). Payloads that are not a multiple of 160 bytes may cause audio gaps.

​
Stream lifecycle

Client → Vendor:  connected → start → media → stop
Vendor → Client:                       media → mark → clear
1

Connection established

A WebSocket connection is opened. The system sends a connected event as the initial handshake.
2

Stream initialized

A start event is sent with stream metadata, caller information, and audio format details.
3

Audio streaming

media events are exchanged bidirectionally — the system sends caller audio, the vendor returns processed audio for playback.
4

Stream terminated

A stop event is sent when the call ends. clear events can interrupt buffered audio at any time.

​
Events sent to vendor

​
Connected

First handshake message after WebSocket connection is established. 
{
  "event": "connected"
}

​
Start

Contains stream metadata. Sent once at stream initialization. 
{
  "event": "start",
  "sequenceNumber": "1",
  "start": {
    "accountSid": "ACXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX",
    "streamSid": "MZXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX",
    "callSid": "CAXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX",
    "from": "XXXXXXXXXX",
    "to": "XXXXXXXXXX",
    "direction": "outbound",
    "mediaFormat": {
      "encoding": "audio/x-mulaw",
      "sampleRate": 8000,
      "bitRate": 64,
      "bitDepth": 8
    },
    "customParameters": {
      "FirstName": "Jane",
      "LastName": "Doe",
      "RemoteParty": "Bob"
    }
  },
  "streamSid": "MZXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX"
}
FieldTypeDescription
eventstringAlways "start"
sequenceNumberstringMessage order tracking (starts at 1)
start.streamSidstringUnique stream identifier
start.accountSidstringAccount identifier
start.callSidstringCall identifier
fromstringOriginating phone number
tostringDestination phone number
start.directionstring"inbound" or "outbound"
start.mediaFormatobjectAudio format specifications
start.customParametersobjectCustom key-value data passed with the stream

​
Media

Encapsulates raw audio data from the caller. 
{
  "event": "media",
  "sequenceNumber": "3",
  "media": {
    "chunk": "1",
    "timestamp": "5",
    "payload": "no+JhoaJjpz..."
  },
  "streamSid": "MZXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX"
}
FieldTypeDescription
media.chunkstringSequential chunk number (starts at 1)
media.timestampstringPresentation timestamp in milliseconds from stream start
media.payloadstringBase64-encoded raw audio

​
Stop

Indicates stream termination or call end. 
{
  "event": "stop",
  "sequenceNumber": "5",
  "stop": {
    "accountSid": "ACXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX",
    "callSid": "CAXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX",
    "reason": "The caller disconnected the call"
  },
  "streamSid": "MZXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX"
}
FieldTypeDescription
stop.accountSidstringAccount identifier
stop.callSidstringCall identifier
stop.reasonstringReason for stream termination

​
DTMF

Sent when touch-tone digits are detected in the inbound audio stream. 
{
  "event": "dtmf",
  "streamSid": "MZXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX",
  "sequenceNumber": "5",
  "dtmf": {
    "digit": "1"
  }
}
FieldTypeDescription
dtmf.digitstringDetected digit (0–9, *, #)

​
Mark

Indicates completion of media playback. Used to synchronize audio delivery between client and vendor. 
{
  "event": "mark",
  "sequenceNumber": "4",
  "streamSid": "MZXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX",
  "mark": {
    "name": "mark label"
  }
}
FieldTypeDescription
mark.namestringCustom label for tracking playback completion

​
Events received from vendor

​
Media

The vendor sends audio data for playback to the caller. The payload must be audio/x-mulaw at 8000 Hz, Base64-encoded PCM mono. 
{
  "event": "media",
  "streamSid": "MZXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX",
  "media": {
    "payload": "a3242sa...",
    "chunk": 1
  }
}

​
Mark

The vendor signals that sent media has finished playing. 
{
  "event": "mark",
  "streamSid": "MZXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX",
  "mark": {
    "name": "my label"
  }
}

​
Clear

Interrupts any buffered audio and resets the stream state. Use this to stop playback immediately (e.g., when the caller interrupts). 
{
  "event": "clear",
  "streamSid": "MZXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX"
}

​
Implementation notes

The streamSid must be unique and consistent across all events for a given stream. Use the value from the start event for all subsequent messages.
  • Base64 encoding ensures compatibility and reliable transmission across systems
  • mark events synchronize playback completion between endpoints — use them to know when audio has finished playing before sending the next segment
  • clear events allow you to interrupt the current audio (e.g., for barge-in when the caller speaks over the bot)
  • Media messages are sent every 100ms; ensure your endpoint can process them at this rate
  • sequenceNumber tracks message order and increments with each event sent