Skip to content
Get packs

Guide · Open source

How to use Buzz: transcribe and translate audio offline

Updated · written with AI from the project's own documentation and checked against our catalogue data

Buzz is a free, open-source app that transcribes and translates audio on your own computer, using OpenAI's Whisper speech recognition model. This guide shows you how to install it, transcribe a file or YouTube link, and export transcripts as TXT, SRT or VTT. It is written for non-experts, so no programming is needed unless you choose the PyPI route.

Buzzchidiwilliams/buzz · ★ 22.4k · MIT

Quick start

  1. 1Pick your install route. Choose the download for macOS or Windows from SourceForge, Flatpak or Snap on Linux, or PyPI if you have Python 3.12.
  2. 2Install prerequisites if needed. For the PyPI route only, install ffmpeg and use a Python 3.12 environment.
  3. 3Launch Buzz. Open the installed app, or run python -m buzz after a PyPI install.
  4. 4Enable GPU acceleration if available. With an Nvidia GPU on version 1.4.6 or later, install CUDA Acceleration from Help, then About Buzz.
  5. 5Import your media. Add an audio or video file or a YouTube link to transcribe.
  6. 6Check the transcript. Use the viewer to search the text and adjust playback speed while you correct it.
  7. 7Export the result. Save the transcript as TXT, SRT or VTT.

Choose how to install Buzz on your computer

The README lists a separate installation path for each operating system. On macOS and Windows you download the installer from SourceForge. On Linux you can use Flatpak, Snap or an AppImage. The PyPI route is an alternative for any system where you are comfortable running Python commands.

  • macOS: download the .dmg from SourceForge. Buzz now requires Apple silicon.
  • Windows: get the installation files from SourceForge. The app is not signed, so you will see a warning. Select More info, then Run anyway.
  • Linux: install with Flatpak, Snap or AppImage.
  • Any system with Python 3.12: install through PyPI, which also requires ffmpeg.

Install Buzz on Linux with Flatpak or Snap

For Flatpak, the README gives this command. Flatpak is the simplest Linux route shown in the source.

For Snap, the README first installs some audio libraries and then installs the snap package.

flatpak install flathub io.github.chidiwilliams.Buzz sudo apt-get install libportaudio2 libcanberra-gtk-module libcanberra-gtk3-module sudo snap install buzz

Install Buzz with PyPI and run it

Before using PyPI, install ffmpeg and make sure you are using a Python 3.12 environment. Then install the package and start Buzz as a Python module. The README also notes separate steps for Nvidia GPU support on Windows with PyPI, using CUDA-enabled PyTorch.

  • Install ffmpeg from the official ffmpeg download page.
  • Use a Python 3.12 environment.
  • For Nvidia GPU support on Windows with PyPI, follow the README's CUDA torch instructions.
pip install buzz-captions python -m buzz

Enable Nvidia CUDA acceleration

If you have an Nvidia GPU and run Buzz version 1.4.6 or later, the README says to open Help, then About Buzz, and install CUDA Acceleration. Buzz also supports Apple Silicon Macs and Vulkan acceleration for Whisper.cpp on most GPUs, including integrated GPUs.

  • Nvidia: Help, then About Buzz, then install CUDA Acceleration (version 1.4.6 or later).
  • Apple Silicon: supported by the app.
  • Other GPUs: Vulkan acceleration for Whisper.cpp is listed as a feature.

Transcribe a file, a YouTube link or a microphone

Buzz can transcribe audio and video files, as well as YouTube links. For events or presentations, it can also transcribe live from a microphone, and a presentation window can display the text for an audience. The README also lists speech separation before transcription, which it says can improve accuracy on noisy audio, and speaker identification in transcribed media.

  • Import an audio or video file, or paste a YouTube link.
  • For live use, choose microphone transcription and optionally open the presentation window.
  • Turn on speech separation if the recording is noisy.
  • Use speaker identification when you need to know who is speaking.

Review and export your transcript

The Advanced Transcription Viewer offers search, playback controls and speed adjustment. You can export transcripts to TXT, SRT and VTT. SRT and VTT are subtitle formats, so they suit video captions. Keyboard shortcuts are available for faster navigation.

  • Search within the transcript.
  • Adjust playback speed while checking the text.
  • Export as TXT for plain text, or SRT or VTT for subtitles.

Use the command line or a watch folder

Buzz includes a Command-Line Interface for scripting and automation, and a watch folder that transcribes new files automatically. A plugin system adds features such as AI summary generation and automated transcript resizing. The README does not list the CLI commands, so check the documentation for details.

  • Watch folder: transcribes new files as they arrive.
  • Command-Line Interface: for scripting and automation.
  • Plugins: for example AI summary generation and transcript resizing.

What people use it for

  • Making a text transcript of an interview or lecture recording without uploading it to a cloud service.
  • Creating SRT or VTT subtitle files for a video.
  • Transcribing a YouTube link.
  • Showing live captions on a presentation window during an event, using a microphone.
  • Automatically transcribing new recordings placed in a watch folder.

Frequently asked questions

Is Buzz free and open source?

Buzz is released under the MIT licence, according to the README badge and project information. The source code is public on GitHub. Check the licence file in the repository for the full terms.

Does Buzz work offline?

Yes. Buzz describes itself as transcribing and translating audio offline on your personal computer, so your files are not sent to a cloud service.

Can I use Buzz on an Intel Mac?

The README says Buzz now requires Apple silicon. The last version to support Intel Macs is 1.4.5, so Intel Mac users should use that version if they need Buzz.

How do I install Buzz on Windows, and why does it show a warning?

Download the installation files from SourceForge. The README says the app is not signed, so you will see a warning. Select More info, then Run anyway.

What do I need to install Buzz with PyPI?

You need ffmpeg and a Python 3.12 environment. Then run pip install buzz-captions and python -m buzz. These prerequisites apply only to the PyPI route.

How does Buzz compare to whisper.cpp, faster-whisper, WhisperX or Subtitle Edit?

The source does not compare these projects, so this guide cannot rank them. Buzz is a desktop app with a graphical interface and features such as live microphone transcription. Compare the tools against your own needs, such as platform, GPU support and subtitle export.

New agent skills, repos and AI tools — every week

The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.

Our verdict

Buzz suits people who want to transcribe or translate recordings offline with a desktop app, especially creators who need SRT, VTT or TXT output. It is a poor fit for Intel Mac users on newer versions, people who need a hosted web service, or anyone who wants a signed Windows installer.

Install options & downloads