Vatis Transcribe icon

Vatis Transcribe

Interact with the Vatis Transcribe API

Actions3

Transcribe → Create Transcript

AI-generated

Summary

Create a transcript from an audio or video file via the Vatis Transcribe API, supporting input as a public URL or binary data from a previous node.

Inputs

  • Media Source (required) — Choose the media input type: a public URL or a binary file from a previous node containing audio or video.
  • Media URL (required) — Public HTTP or HTTPS URL to the audio or video file to be transcribed. Required if Media Source is 'Public URL'.
  • Binary Property (required) — Name of the binary property holding the media file from a previous node. Required if Media Source is 'File (Binary)'.
  • Audio Language — One or more languages of the audio to target during transcription. Leave empty or select 'Auto Detect' to enable automatic language detection. Optionally enable a dedicated Romanian model for Romanian audio.
  • Voice Activity Detection — Settings to enable or configure voice activity detection, including threshold to differentiate speech segments from silence/noise.
  • Diarization — Configure speaker or channel diarization mode, with options for no diarization, speakers (assign text to speakers), or channel (split stereo channels). Speaker counts can also be specified.
  • Audio Intelligence — Enable and configure enhanced transcription, sentiment analysis, transcript summarization with tone and format options, and up to three custom prompts for audio intelligence queries and JSON output formatting.

Output shape

a list of items each containing a JSON object with streamId, streamConfigurationTemplateId, and API uploadResponse details

Each output item corresponds to the input item and includes the unique stream ID generated for the transcript request, configuration info, and the raw upload response from the Vatis API. Errors in individual items return error info if 'Continue on Fail' is enabled.

Examples

Example 1: Transcribe audio from a public URL in English with automatic language detection, voice activity detection enabled, and no diarization.

Set Media Source to 'Public URL', provide the Media URL, leave Audio Language empty, enable Voice Activity Detection (default settings), set Diarization Mode to 'No Diarization'. Expect output with a streamId identifying this transcription job.

Example 2: Transcribe a binary audio file from a previous node with speaker diarization for up to 3 speakers, sentiment analysis, and summary generation.

Set Media Source to 'File (Binary)', specify the Binary Property name, set Diarization Mode to 'Speakers' with maximum 3 speakers, enable Sentiment Analysis and Summary in Audio Intelligence settings. Output includes streamId and details of the initiated transcription.

Links

Discussion