Skip to main content

Create Transcription Node

The Create Transcription node transcribes audio files to text using OpenAI's Whisper model.

Create Transcription Node

Description​

The Create Transcription node converts speech in an audio file to text using OpenAI's Whisper transcription model. It supports various audio formats and can transcribe in multiple languages.

Execution Inputs​

  • Execute (execution): Triggers the transcription request

Variable Inputs​

  • AudioPath (string): Path to the audio file to transcribe
  • Model (string, optional): Model to use. Must be "whisper-1" (default: "whisper-1")
  • Prompt (string, optional): Optional text prompt to guide the transcription (e.g., proper nouns, technical terms)
  • ResponseFormat (string, optional): Format of the response. Must be "json", "text", "srt", "verbose_json", or "vtt" (default: "json")
  • Temperature (float, optional): Sampling temperature between 0 and 1 (default: 0)
  • Language (string, optional): Language code (e.g., "en", "es", "fr"). If not specified, language is auto-detected

Execution Outputs​

  • Success (execution): Execution continues here if the request succeeded
  • Failed (execution): Execution continues here if the request failed

Variable Outputs​

  • Response (string): Full JSON response from the API
  • Text (string): The transcribed text
  • ErrorMessage (string): Error message if request failed

Special Notes​

  • Requires a valid API key configured via Set API Key node
  • Audio file must be in a supported format (mp3, mp4, mpeg, mpga, m4a, wav, webm)
  • Maximum file size is 25MB
  • The Prompt can help with proper nouns, technical terms, or context
  • Language auto-detection works well but specifying language can improve accuracy
  • ResponseFormat "text" returns plain text, "json" returns structured data

Example Usage​

Transcribe an audio file:

  1. Set AudioPath to "audio/recording.mp3"
  2. Set Model to "whisper-1"
  3. Set ResponseFormat to "text"
  4. Connect Execute to transcribe
  5. Use Text output to get the transcription

Transcribe with language and prompt:

  1. Set AudioPath to "audio/meeting.mp3"
  2. Set Language to "en"
  3. Set Prompt to "John Smith, Sarah Johnson, Q4 earnings"
  4. Connect Execute to transcribe

Flow Download Example​

Placeholder for flow download example

Additional Pictures​

Placeholder for additional node images and examples