Deepgram Transcriber icon

Deepgram Transcriber

Transcribes audio using Deepgram API (pre-recorded)
Join our community

Overview

This node transcribes audio files using the Deepgram API. It supports transcription from either a publicly accessible audio URL or an audio file provided as binary data from a previous node. Users can select from various transcription models tailored to different audio types and specify additional transcription options such as language, punctuation, diarization, and keyword boosting. The node outputs either the full raw transcription response or just the transcript text, optionally including metadata about the transcription process. This node is useful for automating transcription workflows in scenarios like meeting recordings, phone calls, video content, or financial audio analysis.

Use Case Examples

  1. Transcribing a podcast episode from a public URL to generate searchable text.
  2. Transcribing customer support call recordings stored as binary files in n8n for quality analysis.
  3. Automatically transcribing meeting recordings with speaker diarization enabled to identify different speakers.

Properties

Name Meaning
Input Source Type Select whether the audio input is a URL or a binary file from a previous node.
Audio URL The publicly accessible URL of the audio file to transcribe. Required if Input Source Type is URL.
Binary Property The name of the binary property containing the audio file data. Required if Input Source Type is Binary File.
Model Choose the transcription model that best fits the audio type, such as Base, Enhanced, Nova 2 variants, or Nova 3 (latest).
Additional Options Additional transcription parameters such as callback URL, diarization, keywords boosting, language, punctuation, and smart formatting.
Append Metadata Whether to include metadata such as endpoint used, parameters, and transcription duration in the output.
Output Format Choose to output either the full raw transcription response or only the extracted transcript text.
Transcript Field Name Name of the JSON field to store the transcript text when 'Transcript Only' output format is selected.

Output

JSON

  • deepgramTranscription - The full transcription response object from Deepgram API when 'Full Raw Transcript' output format is selected.
    - The extracted transcript text stored under the user-defined field name when 'Transcript Only' output format is selected.
  • deepgramMetadata - Optional metadata including API endpoint used, parameters sent, and transcription duration in milliseconds, included if Append Metadata is enabled.

Dependencies

  • Deepgram API key credential required for authentication.

Troubleshooting

  • Ensure the audio URL is publicly accessible if using URL input source; otherwise, the transcription will fail.
  • Verify the binary property name matches the actual binary data property when using binary file input source.
  • Check that the Deepgram API key credential is correctly configured and has necessary permissions.
  • If the transcription fails, error messages from Deepgram API are propagated; review these messages for issues like unsupported audio format or invalid parameters.
  • When using asynchronous callback URL, ensure the URL is reachable and correctly handles incoming webhook data.

Links

Discussion