Audio & Realtime

Audio workloads split into two shapes. Use the audio endpoints for file-like requests such as text-to-speech, transcription, and audio translation. Use the realtime WebSocket endpoint when the user experience needs low-latency, interactive audio or multimodal events.

Choose The Workflow

Workflow	Endpoint	Use it when
Text to speech	`POST /v1/audio/speech`	You need an audio file from text.
Transcription	`POST /v1/audio/transcriptions`	You need text from an audio file.
Audio translation	`POST /v1/audio/translations`	You need translated text from an audio file.
Realtime WebSocket	`WS /v1/realtime?model={model}`	You need bidirectional streaming audio or realtime multimodal events over WebSocket. Plain `GET /v1/realtime` only returns metadata.

Discover Models

Query the model catalog before hard-coding a model. Use recommended shortlists for speech and transcription, and use model details to confirm realtime support before opening a socket.

curl "https://api.aisonar.dev/v1/models?recommended_for=tts" \
  -H "Authorization: Bearer sk-your-api-key"

curl "https://api.aisonar.dev/v1/models?recommended_for=stt" \
  -H "Authorization: Bearer sk-your-api-key"

Synchronous Audio Requests

Speech, transcription, and translation requests return directly from the HTTP request. Large inputs can take longer than common client defaults, so set a generous timeout and store request IDs for support.

curl -X POST "https://api.aisonar.dev/v1/audio/speech" \
  -H "Authorization: Bearer sk-your-api-key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "tts-1-hd",
    "voice": "nova",
    "input": "Welcome to AI Sonar."
  }' \
  --output speech.mp3

Realtime Sessions

Open a WebSocket with the model in the query string and the API key in the Authorization header. Keep the event format documented for the selected realtime model, and close the socket when the session is complete. This guide covers the WebSocket proxy only. AI Sonar does not currently provide OpenAI Realtime REST client secret, translation client secret, Calls, or legacy beta session management endpoints.

import WebSocket from 'ws';

const socket = new WebSocket('wss://api.aisonar.dev/v1/realtime?model=gpt-realtime', {
  headers: { Authorization: 'Bearer sk-your-api-key' }
});

socket.on('message', (event) => console.log(event.toString()));

State Handling

Save generated audio files instead of replaying the same request on refresh.
For transcription and translation, show upload and processing states even when the API call is synchronous.
For realtime, handle close events and reconnect only after the user starts a new session.
Do not put API keys, private URLs, or account secrets in audio text input.

API Reference

Topic	Reference
Create Speech	Create Speech
Create Transcription	Create Transcription
Create Translation	Create Translation
Realtime WebSocket	Realtime WebSocket
List Models	List Models
Billing & Pricing	Billing & Pricing

Getting Started

Core Guides

Media Guides

Choose The Workflow

Discover Models

Synchronous Audio Requests

Realtime Sessions

State Handling

API Reference

​Choose The Workflow

​Discover Models

​Synchronous Audio Requests

​Realtime Sessions

​State Handling

​API Reference

Choose The Workflow

Discover Models

Synchronous Audio Requests

Realtime Sessions

State Handling

API Reference