Audio

Prepare a direct Spectra-2 stream

Server-only setup for one recording in an existing prepared session. Open the returned URL with [protocol, ticket] subprotocols, then use the documented start, PCM16, finish and cancel messages. Finish must precede the original deadline. The result matches the batch schema. No automatic reconnect or inference retry.

post/v1/audio/stream-sessions

Request

  • Base URL: https://speech-api.oruk.ai
  • URL: https://speech-api.oruk.ai/v1/audio/stream-sessions
  • Auth: HTTP bearer

Headers

X-Oruk-Sessionstring required

The existing prepared-session token.

X-Request-IDstring required

One unused request ID supplied by that session.

Request body

model'oruk-spectra-2' required
task'transcription' | 'analysis' required

Response

Prepared connection; no inference has run.

object'speech.stream_session' required
model'oruk-spectra-2' required
task'transcription' | 'analysis' required
request_idstring required
urlstring uri required

Open this WebSocket URL from your server. Do not add query parameters.

protocol'oruk.spectra2.v1' required
ticketstring required

Private one-request credential. Send [protocol, ticket] as WebSocket subprotocols; never log it.

expires_at_msinteger required

Original prepared-session deadline, as Unix milliseconds.

max_audio_secondsnumber required

Time remaining at setup. Honor the fresher recording limit in the WebSocket ready message.

Changes

No recorded changes to this operation across all 1 revision of this API.