Skip to main content

Transcribe Speech Node

The Transcribe Speech node converts audio input into text using a selected OpenAI Whisper model and reference text. This powerful node enables accurate speech-to-text conversion, making it ideal for transcription services, voice input processing, accessibility features, and language learning applications.

Transcribe Speech node

Basic Usage​

Use Speech Input Node, Display Text, Text, and Transcribe Speech Node for your process.

Inputs​

The Transcribe Speech node accepts the following inputs:

  • Input: The audio file or recording to be transcribed into text.
  • Reference (Optional): Reference text that can help improve transcription accuracy, especially for specialized vocabulary, names, or technical terms.

Outputs​

  • Output: The transcribed text from the audio input.

Configuration​

Model Selection​

Select the OpenAI Whisper model to use for transcription:

  • OpenAI Whisper: Advanced speech recognition model supporting multiple languages and accents
    • High accuracy for various audio qualities
    • Supports 99+ languages
    • Handles background noise and multiple speakers
    • Automatic language detection

Available Models​

  • OpenAI Text to Speech: General-purpose text-to-speech model by OpenAI.
  • OpenAI Alloy: A distinct voice option from OpenAI.
  • OpenAI Fable: A voice option from OpenAI designed for storytelling.
  • OpenAI Nova: A unique voice option from OpenAI.
  • OpenAI Onyx: A further voice option from OpenAI.
  • 11LAB Sue: A high-fidelity voice model from 11LABS.
  • 11LAB Katie: Another voice model from 11LABS.
  • 11LAB Alexander: A third voice model from 11LABS.

Usage​

Setting Up Transcribe Speech​

  1. Add the node to your flow canvas.
  2. Select Model: Choose OpenAI Whisper for speech transcription.
  3. Connect Input: Link from upstream nodes (e.g., Speech Input Node, File Upload) to provide the audio content.
  4. Connect Reference (Optional): Link reference text to improve accuracy for specialized content.
  5. Connect Output: Link to downstream nodes (e.g., Display Text, Text Join, AI General Prompt) to use the transcribed text.

Example Workflows​

Convert audio input into Note Transcription​

Scenario: Create a voice note transcription tool that converts spoken recordings into written text for easy review and editing.

Transcribe Speech Example

Steps to Create the Flow:

  1. Add a Start Node.

  2. Add and connect a Speech Input Node for audio recording.

    • Configure to allow voice recording
    • Set appropriate recording duration limits

  3. Add and connect a Transcribe Speech Node.

    • Select Model: OpenAI Whisper
    • Connect the Speech Input Node's Output to Input

  4. Add and connect a Display Text node to show the transcribed text.

Result: Users can record voice notes and immediately see them transcribed into text, making it easy to capture thoughts and ideas verbally.