Skip to main content
Cawme supports four call types, each suited for different use cases. Set the type field when creating a call.

​
FIXED_TTS

Convert text to speech and play it during the call. The simplest call type — just provide a message and a voice. 
{
  "instanceId": "inst_abc123",
  "type": "FIXED_TTS",
  "target": { "phoneNumber": "+14155550123" },
  "tts": {
    "text": "Hi John, your appointment is confirmed for tomorrow at 2pm.",
    "voiceId": "en-US-001"
  }
}
Use cases: Order notifications, appointment reminders, payment alerts, one-way announcements.

​
FIXED_AUDIO

Play a pre-recorded audio file (Base64 encoded MP3 or WAV). Best when you need a polished, branded voice message. 
{
  "instanceId": "inst_abc123",
  "type": "FIXED_AUDIO",
  "target": { "phoneNumber": "+14155550123" },
  "audio": {
    "audioBase64": "SUQzBAAAAAAA..."
  }
}
Use cases: Marketing campaigns, branded announcements, compliance recordings.

​
DYNAMIC_STREAM

Stream real-time PCM or MP3 audio frames over a WebSocket. Use this when you want full control over the audio pipeline — connect your own TTS, translation layer, or live human agent. 
{
  "instanceId": "inst_abc123",
  "type": "DYNAMIC_STREAM",
  "target": { "phoneNumber": "+14155550123" }
}
After the call is created, establish a WebSocket connection at GET /v1/calls/:callId/stream:
  • Client → Server: Send PCM or MP3 audio frames
  • Server → Client: Receive real-time transcripts and call events
Use cases: Live agent-assisted calls, custom voice bots, real-time translation.
WebSocket streaming is currently in preview. See Stream endpoint for details.

​
DYNAMIC_AGENT

Deploy an AI agent that conducts the call autonomously based on natural-language instructions. The agent handles the conversation flow without any manual input. 
{
  "instanceId": "inst_abc123",
  "type": "DYNAMIC_AGENT",
  "target": { "phoneNumber": "+14155550123" },
  "agent": {
    "instructions": [
      "Greet the person politely",
      "Ask if they are interested in a product demo",
      "If yes, collect their preferred date and time",
      "Confirm the appointment before ending the call"
    ]
  }
}
Use cases: Sales prospecting, appointment scheduling, customer surveys, lead qualification.

​
Comparison

FeatureFIXED_TTSFIXED_AUDIODYNAMIC_STREAMDYNAMIC_AGENT
Pre-written message✓✓——
Custom audio file—✓——
Real-time audio control——✓—
AI-driven conversation———✓
Requires WebSocket——✓—

Start a call

POST /v1/calls — full request reference