Realtime API

WebSocket Endpoint

Connect via WebSocket for realtime communication. Supports OpenAI Realtime API protocol.

Query Parameters

• session_id - Connect to existing session
• model - Auto-create session with model (OpenAI-compatible)

Authentication

Browser WebSocket can't set headers, so send auth as first message:

// Browser JavaScript
ws.onopen = () => {
  ws.send(JSON.stringify({
    type: 'auth',
    api_key: 'rtav_ak_your_api_key_here'
  }));
};

Examples

// Browser JavaScript
// Auto-create session (OpenAI-compatible)
const ws1 = new WebSocket('wss://api.rtav.io/v1/realtime?model=gpt-5.2');

// Connect to existing session
const ws2 = new WebSocket('wss://api.rtav.io/v1/realtime?session_id=session_abc123');

See WebSocket Guide for complete examples.

WebRTC Endpoint

Create a WebRTC call for low-latency audio/video streaming.

Request

Multipart form data with SDP offer and session configuration:

• sdp - SDP offer from client
• session - Session configuration (JSON string)

Authentication

Use ephemeral key (client secret) obtained from /v1/realtime/client_secrets:

Authorization: Bearer ek_68af296e8e408191a1120ab6383263c2

Response

Returns SDP answer as plain text:

v=0
o=- 1234567890 1234567890 IN IP4 0.0.0.0
s=-
t=0 0
...

See Create Call Reference and WebRTC Guide for complete examples.

Client Secrets (Ephemeral Keys)

Get ephemeral keys for WebRTC calls (OpenAI-compatible):

Request Body

// Browser JS / Node.js
const response = await fetch('https://api.rtav.io/v1/realtime/client_secrets', {
  method: 'POST',
  headers: {
    'Authorization': 'Bearer rtav_ak_your_api_key',
    'Content-Type': 'application/json'
  },
  body: JSON.stringify({
    session: {
      type: 'realtime'
    }
  })
});

Response

{
  "client_secret": "ek_68af296e8e408191a1120ab6383263c2",
  "expires_in": 60
}

Event Types

Both WebSocket and WebRTC (data channel) support the same event types following the OpenAI Realtime API protocol:

Client Events

• auth - Authenticate with API key
• session.update - Update session configuration
• conversation.item.create - Create conversation item (text message)
• input_audio_buffer.append - Append audio data (WebSocket only)
• input_audio_buffer.commit - Commit audio buffer (WebSocket only)
• response.create - Trigger response generation

Server Events

• session.created - Session created
• session.updated - Session updated
• response.created - Response generation started
• response.output_audio.delta - Audio output chunk (WebSocket: base64, WebRTC: RTP)
• response.output_image.delta - Video frame (WebSocket: base64 JPEG, WebRTC: RTP)
• response.output_audio_transcript.delta - Output transcript chunk
• conversation.item.input_audio_transcription.completed - Input audio transcription completed
• response.done - Response generation completed