AssemblyAI icon

AssemblyAI

Transcribe audio and video files using AssemblyAI's speech-to-text and speech understanding AI models.

Actions12

Transcript → Create

AI-generated

Overview

This node creates a new transcription of an audio file using AssemblyAI's speech-to-text and speech understanding AI models. It is useful for converting spoken audio content into text, enabling further analysis, search, or processing. For example, it can transcribe interviews, meetings, podcasts, or any audio recordings by providing the audio file URL and optional advanced transcription settings.

Use Case Examples

  1. Transcribing a podcast episode by providing its audio URL and enabling automatic highlights to capture key phrases.
  2. Creating a medical transcription by setting the domain to 'Medical' and adding custom spelling rules for medical terminology.
  3. Using language detection and multichannel transcription to transcribe a multilingual audio file with multiple speakers.

Properties

Name Meaning
Audio URL URL of the audio file to transcribe, required for creating a transcription.
Additional Fields Optional advanced transcription settings such as start/end times, auto highlights, content safety, custom spelling, language options, speaker labels, profanity filtering, punctuation, redaction of personally identifiable information (PII), summarization (deprecated), speech models, webhook settings, and more to customize the transcription process.

Output

JSON

  • json
    • id - Unique identifier of the created transcription.
    • status - Current status of the transcription (e.g., queued, processing, completed).
    • audio_url - URL of the audio file being transcribed.
    • text - The transcribed text output.
    • words - Array of words with timestamps and confidence scores.
    • chapters - Automatically generated chapters if enabled.
    • highlights - Automatically highlighted key phrases if enabled.
    • content_safety_labels - Detected sensitive content labels if content safety is enabled.
    • speaker_labels - Information about identified speakers if speaker labels are enabled.
    • redacted_text - Transcript text with personally identifiable information redacted if enabled.
    • webhook_url - Webhook URL to notify when transcription is complete.

Dependencies

  • AssemblyAI API key credential

Troubleshooting

  • Ensure the provided Audio URL is accessible and publicly reachable by AssemblyAI API.
  • If using advanced options like language detection or speaker labels, verify that the account supports these features.
  • Check that required fields like Audio URL are not empty to avoid validation errors.
  • For redaction features, ensure the correct PII policies and options are set to avoid incomplete redaction.
  • If the node returns API errors, inspect the error message for details such as invalid parameters or authentication issues and adjust accordingly.

Links

Discussion