Actions14
Transcription → Create and Wait [Deprecated] → Transcribe
AI-generatedSummary
Creates a Soniox transcription from an existing file, polls until it completes or fails, then retrieves the transcript. This deprecated compatibility operation is retained for existing workflows; use Transcribe or Create Job for new workflows.
Inputs
- File ID (required) — Soniox file ID to transcribe; the implementation requires a UUID.
- Model Name or ID (required) — Transcription model, loaded from Soniox or supplied as an ID; defaults to stt-async-v5.
- Additional Fields — Optional context (general JSON, terms, free-form text, translation terms), language identification or hints, speaker diarization, and translation settings. These guide the transcription request.
- Options — Check Interval (default 5 seconds) controls polling frequency; Max Wait Time (default 300 seconds) sets the timeout. The node polls immediately before waiting between subsequent checks.
Output shape
One n8n item per input item, containing the transcription status response fields plus a top-level text field with the retrieved transcript.
If speaker diarization is enabled and the transcript includes tokens, output also includes speaker segments with speaker, text, and start/end timestamps. The operation does not include token-level data in its output. It fails on a terminal transcription error or timeout; optional cleanup of the transcription is not exposed in this operation's UI.