Unable to generate transcript from MP3

ChatGPT
1

I just paid for gpt-4 so that I could generate transcripts from MP3 files. I immediately got this cryptic error message:
“The required module for transcription is not available in the current environment”
How do I solve this problem?

2

Hi @earthmanrobert !
Welcome :people_hugging: to the community.

I think you subscribed ChatGPT Plus.
If its so, it cannot transcribe audio files to text.

You need to use Wisper model.
You should read read following link on OpenAI docs:

Speech to Text

The Audio API provides two speech to text endpoints, transcriptions and translations, based on our state-of-the-art open source large-v2 Whisper model. They can be used to:

  • Transcribe audio into whatever language the audio is in.
  • Translate and transcribe the audio into english.

File uploads are currently limited to 25 MB and the following input file types are supported: mp3, mp4, mpeg, mpga, m4a, wav, and webm.

3

Thank you for your reply. I am sorry, but I do not understand at all. A Google search shows many hits confirming that ChatGPT can transcribe audio files.

I never heard of this Whisper thing. Do I have to pay separately for that? How do I use it? Do I need to pay for ChatGPT to use it? Thank you again for your help. The “help” at OpenAI is utterly useless. Thanks!

4

When you search on Google, you might find information suggesting that ChatGPT can transcribe audio files. While Google search might show general information about ChatGPT, to avoid confusion, always refer to official OpenAI sources for the most accurate guidance. Here are some trusted places to learn more:

  1. OpenAI Documentation
  2. API Reference
  3. OpenAI Pricing
  4. OpenAI Blog
  5. OpenAI Help-FAQ

Now, let’s talk about transcription. Transcription is not done directly within the ChatGPT interface. ChatGPT free or ChatGPT Plus do not directly support audio transcription (for now). Instead, OpenAI offers a separate service for this purpose called Whisper, which you access via the Audio API.

If you ask “Do I need ChatGPT Plus to use the API?”, answer is “NO!”

You don’t need a ChatGPT Plus subscription to use the Whisper model or any API service.

  • API Services Are Separate: The Whisper model is accessed via the Audio API, which is a separate service from the ChatGPT interface.
  • Pay for What You Use: When you use the API, you’re billed based on how much you use it (e.g., the amount of audio transcribed or tokens processed), not through a subscription. This is a pay-as-you-go system.

To use Whisper, you need an OpenAI Account:

  1. Create an OpenAI Account: If you don’t have one yet, sign up at Platform OpenAI’s website.
  2. Create an API Key: This key will allow you to use Whisper and other API services.
  3. Follow the Speech to Text Guide: This explains how to transcribe audio using the Whisper model.

Pricing details for audio models
Pricing details for audio models869×515 15.2 KB

HTH @earthmanrobert

1 Like
5

Thank you for that answer. I now know that this Whisper thing is not at all user friendly. I have no idea how to deal with an API. I am guessing most people are in the same situation.

And that it is too limited in capacity for what I need. 25MB seems quite small.

It seems bizarre that so much development went into the transcription engine and nothing went into making it usable for normal people.

2 Likes
6

hat the same idea like you, just half a year before. battled with support, till the finnaly “got it”, tldr. Aint possible with app or web… I switched to ai studio and gemini 2 a few days ago. free. done but no data control.

the usual stuff you get from community is the “use whipser” answer, break ya finers with python and , which, ayou you found also doent help you.

BUT, if you want to stick with openai, use this : https://platform.openai.com/playground/chat?models=gpt-4o-audio-preview
feed it with 5$ credit and have fun.
but be carefull, it will cost you. like 3. cents a minute or so :wink: