Skip to main content
Chatbot Builder AI integrates with OpenAI’s Whisper model to convert audio files sent by users (voice notes or uploads) into text. This allows you to process voice commands, save verbal feedback as text, or simply reply to voice messages with AI.

​
Prerequisites

You must have an active OpenAI Integration connected to your account. See the OpenAI Connection Guide if you haven’t set this up yet.

​
Configuration Flow

1

Collect User Audio

First, you need to capture the audio file from the user.
  1. Add a Get User Data block to your flow.
  2. Set the Reply Type to Audio (or File/Any depending on your specific use case).
  3. Save to Custom Field: Create a variable to store the audio URL (e.g., user_audio_file).
  4. Optional: Add a text prompt like “Please record your message now.”
2

Configure OpenAI Action

Add an Action block immediately after the user input.
  1. Select Integrations > OpenAI > Convert Speech to Text.
  2. Audio Input: Select the Custom Field you saved in Step 1 (user_audio_file).
  3. Save Output: Choose a destination Custom Field for the resulting text (e.g., transcribed_text).
3

Use the Text

You can now use the transcribed_text variable just like any other text input.
  • Display it: Add a Send Message block: “I understood: transcribed_text”.
  • Process it: Feed transcribed_text into another AI Action to generate a response to the user’s question.

​
Use Cases

Voice Support

Allow users to explain complex issues via voice note, transcribe it, and send it to your support ticketing system.

Accessibility

Automatically transcribe voice messages for users who prefer reading or for internal logging purposes.