realtime.response.audio.delta

Streaming audio chunks (optional)

Triggered for each audio chunk as the AI generates speech audio. This event is optional and primarily used when receiving audio via WebSocket. For most applications, WebRTC is the recommended method for audio/video streaming. Handle this event only if you’re using WebSocket for audio streaming instead of WebRTC.

Event Properties

type · string

Event type. Always "realtime.response.audio.delta" for this event.

data · object

Event data object containing audio information.

data.delta · string

Base64-encoded audio chunk.

Example: "base64-encoded-audio-chunk"

Example

{
  "type": "realtime.response.audio.delta",
  "data": {
    "delta": "base64-encoded-audio-chunk"
  }
}

Usage Example

const NavTalkMessageType = Object.freeze({
    REALTIME_RESPONSE_AUDIO_DELTA: "realtime.response.audio.delta",
    // ... other event types
});

async function handleReceivedMessage(data) {
    const nav_data = data.data;
    
    switch (data.type) {
        case NavTalkMessageType.REALTIME_RESPONSE_AUDIO_DELTA:
            if (nav_data.delta) {
                // Process audio chunk
                processAudioChunk(nav_data.delta);
            }
            break;
    }
}

For most applications, WebRTC is recommended for audio/video streaming.