Free Audio to Text Transcription Guide Using Audacity

An easy guide for turning voice notes & podcasts into text with Audacity + OpenVINO Whisper

This guide helps you quickly turn audio voice notes, phone calls or podcasts into ready-to-edit text using Audacity and Intel’s OpenVINO Whisper plugin — no coding required. Take ANY MP3, Wav or audio file and extract the spoken word as text.

The best part is it is completely free, and open source, with no payment, subscription or ongoing cost. Below is the install process for Windows, follow the same sort of steps on Mac and “translate” for your operating system. Download the relevant files for your OS and device, links below.

Step 1. Install Audacity

If you are not already using Audacity then you need to install it. Click here:

1. Go to https://www.audacityteam.org/download/
2. Choose your system (Windows/macOS).
3. Follow the installation steps (click “Next” a few times until complete).
4. Open Audacity to check it installed correctly and works.

Once Audacity is installed you are ready to install, configure and run the openVINO Whisper module for Audacity that transcribes. You also have a very powerful audio editing suite, that is absolutely 100% free to use, no subscription, credit card or email address required!

Step 2. Install the OpenVINO Whisper Plugin

This plugin allows Audacity to understand and transcribe speech directly inside the program.

1. Download the official AI Effects plugin from Audacity here: https://www.audacityteam.org/download/openvino/
2. Under the latest release, find the file that matches your version of Audacity (the most recent release should work) ending in `.exe` (for example, `openvino-whisper-win64.exe`).
3. Download it, then double-click the file to install.
4. Follow the setup prompts — it will automatically add the Whisper AI tools to Audacity. It may also prompt you to download additional files. 

You want to download the “small” data set, it’s actually quite big, circa 5GB, so ensure you have the bandwidth and time to download.  The “small” model is best for everyday business use — it’s accurate and usually transcribes 10 minutes of audio in about 1 minute.

5. Once everything is installed for the module you need to activate it within Audacity, so open the program and go to: Edit > Preferences > Import/Export > Modules

Find the module “mod-openvino” and select “enabled” from the drop down alongside the
module’s name.  

6. Finally, close and restart Audacity otherwise the module will not be active and the translation will not work.

Note: If the latest version of Audacity fails to recognise the module once it is installed then roll back to an earlier version of Audacity. Earlier versions are hosted here: https://www.fosshub.com/Audacity-old.html 

Don’t worry if you have to roll-back to use an earlier version, we currently use: audacity-win-3.5.1-64bit and the corresponding plugin for that version.

To find the compatible version of the openVINO module visit Intel’s official plugin development page: https://github.com/intel/openvino-plugins-ai-audacity/releases

Step 3. Open and Prepare Your Audio

Now it is time to get transcribing!

1. Launch **Audacity**.
2. Go to **File > Import > Audio** and choose your MP3 (call, voice note, podcast etc.).
3. (Optional) You can tidy up sound using:

• • Effect > Noise Reduction – to remove background noise.
  • Effect > Normalize – to make the voice clearer.

This is an optional step but might make transcribing your audio more accurate. Always record the best audio you can!

Step 4. Run the Whisper Transcription

1. Once your audio loads, select all the audio on the track you want transcribed.
   Click the “Select” button bottom left of the track controls on the LHS.
2. When the track is selected go to: Analyze > OpenVINO Whisper Translation

In the pop-up window select the following settings:

• • OpenVINO Inference Device: Select “CPU” if it is not preselected.
  • Select: Whisper Model: small (recommended — accurate and fast).
  • Select: Mode: Transcribe
  • Select: Source Language – English (or the language being transcribed)

Click “Apply”.

The audio is transcribed into a track titled “Transcription(small)” beneath the audio track being transcribed. You can transcribe multiple tracks inside one Audacity project, useful for multi-track audio.

Step 5. Save and Use Your Text

We now need to get the text out of Audacity, which you do as follows:

1. Select text transcription track, click “Select” button  (bottom left of track controls on the LHS.)
2. Once selected go to: File > Export Other > Export Labels
3. Choose where you are saving the file and the file name.

Audacity will then export the transcription into a time stamped plain text file, with each line starting with a time stamp followed by the transcription on the same line.

You can then edit the text how you want to format it, making it into a single document, a blog post or allow you to quickly index the information for easy retrieval. 

You can tweak and use this text for:

• Blog posts
• LinkedIn updates
• Instagram captions
• Email newsletters

It can also help when editing video or audio to see which sectons you want to chop outfor a rough edit  before getting into the finer edit details.

Tips for Best Results

A few things to consider through this process

• Record in a quiet space — clear sound gives better accuracy.
• Use your phone’s voice memo app or a USB mic for cleaner audio.
• Shorter the recording the faster the transcription
• Proofread before posting, AI transcription is about 95% accurate with perfect audio and good diction by the person speaking.

We hope that helps you save time and brain space

For the Latest Episodeof the Business Oddcast. Click Here

Human = 100% Edited and ehcked content
A.I. Use = Generate sections and spellcheck