Transcribe audio with free and open-source software on Windows

Whisper.cpp is a program for speech recognition. It can be used for transcribing speech from audio or videos into text. While online services may charge a fee for transcriptions, Whisper.cpp is a free and open-source program that runs on your own computer.

Using a graphical user interface (Vibe)

Vibe is a graphical user interface (GUI) program to make it easier to use Whisper.cpp, which lacks a GUI of its own.

1. Download and run the latest Vibe installer (for Windows, it will be vibe_#.#.#_x64-setup.exe).
2. Launch the program
3. Wait for Vibe to download the base model
4. Click “Select file”, then click “Transcribe”
5. Choose your preferred output format and save the file.

Using the command line (whisper-cli.exe)

1. Download the Whisper.cpp release archive (whisper-bin-x64.zip)
2. Download the speech recognition model (ggml-base.en.bin) from Hugging Face
3. Extract the contents of whisper-bin-x64.zip and navigate to the “Release” sub-folder
4. Create a folder called “models” within the “Release” folder
5. Move the ggml-base.en.bin model you downloaded in step 2 to the “models” subfolder.
6. Open a terminal
7. Navigate to the “Release” folder and execute ./whisper-cli.exe input.wav, where input.wav is the audio file you want to transcribe.