Skip to main content
Live calls let your application start or handle WhatsApp audio calls from a connected token. Version 1 supports audio-only media over WebSocket PCM. WebRTC is planned as a later transport, and the call records already include a transport field so your integration can keep the same call model.

​
Start a call

POST /api/v1/sessions/{sessionId}/calls
Authorization: Bearer <api_key>
Content-Type: application/json

{
  "to": "[email protected]"
}
The token must be paired and connected. The response contains a call snapshot with callId, peerJid, direction, state, transport, timestamps, and mute state.

​
Stream live audio

Upgrade the media URL with bearer auth: 
GET /api/v1/sessions/{sessionId}/calls/{callId}/media
Authorization: Bearer <api_key>
Upgrade: websocket
Client binary frames are outbound mono Float32LE PCM at 16000 Hz. Server binary frames are inbound peer audio in the same format. Server JSON text frames include:
TypePurpose
readyConfirms encoding, sample rate, channels, and direction.
bufferReports live outbound buffer depth and backpressure watermarks.
call_stateMirrors call state changes.
errorReports invalid frames or feed failures.
endedReports that the call ended.
Raw audio is never sent over SSE or webhooks.

​
Play prerecorded audio

Upload an audio asset once, then play it into active calls: 
POST /api/v1/call-audio-assets
Authorization: Bearer <api_key>
Content-Type: multipart/form-data
Multipart field:
FieldRequiredNotes
audioYesAny supported audio/* file up to 32 MB.
Playback: 
POST /api/v1/sessions/{sessionId}/calls/{callId}/audio/play
Authorization: Bearer <api_key>
Content-Type: application/json

{
  "assetId": "019f..."
}

​
Recordings

WADial stores inbound, outbound, and mixed WAV recording artifacts for calls with captured media. Retention defaults to 30 days and can be configured by the account environment. Recording-ready events include metadata such as recordingId, kind, durationMs, sizeBytes, and expiresAt; they do not include raw audio.