Create Transcription Session

POST /v1/realtime/transcription_sessions

Creates an ephemeral token for WebSocket-based real-time streaming transcription.

Request

Content-Type: application/json

curl -X POST "https://api.osmapi.com/v1/realtime/transcription_sessions" \
  -H "Authorization: Bearer $OSM_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"input_audio_transcription": {"model": "whisper-1"}}'

Note: Request body is passed through to OpenAI. See OpenAI documentation for available parameters.

Response

Returns an ephemeral session with a client secret for establishing a WebSocket transcription connection.

Usage

Use the returned token to connect to the Realtime WebSocket with intent=transcription:

const session = await fetch("/v1/realtime/transcription_sessions", { ... });
const { client_secret } = await session.json();
// Connect to wss://api.osmapi.com/v1/realtime?intent=transcription

See the Realtime feature guide for detailed usage.