How to Transcribe Podcasts Offline on Windows

To transcribe a podcast offline on Windows, install a free Whisper app such as Buzz or Vibe and download a model once. Then drag in the episode and export the transcript as text or SRT subtitles. Everything runs on your PC. Two common shortcuts aren’t private, so skip them if privacy is the point: Word’s Transcribe feature stores recordings on OneDrive, and Windows voice typing (Win + H) uses Microsoft’s online speech service.

Which Windows tools transcribe podcasts offline? #

Microsoft’s own documentation confirms the last two. Word stores transcribed recordings in a OneDrive folder, with a limit of 300 minutes of uploaded audio a month for Microsoft 365 subscribers. Voice typing “uses online speech recognition, which is powered by Azure Speech services.”

Step-by-step: transcribe a podcast with Buzz #

Buzz is the simplest free option. It runs OpenAI’s Whisper on your PC and exports TXT, SRT and VTT.

1. Download Buzz from the links on its GitHub page. The project notes that the Windows installer isn’t signed, so expect a SmartScreen warning when you run it.
2. Open the episode file. MP3, WAV and M4A all work, as does video, for video podcasts.
3. Choose a model and the language. Start with a small or medium Whisper model. Set the language to the one spoken on the show rather than auto-detect.
4. Transcribe. Buzz downloads the model the first time. After that it works offline.
5. Review and export. Fix names and numbers in the editor, then export TXT for show notes or SRT/VTT for captions. Buzz can also identify speakers for interview episodes.

Batch a whole back catalog with Vibe #

Vibe is also free and open source. It’s the better choice when you have dozens of episodes. It batch-transcribes multiple files, takes video directly, supports speaker diarization and exports SRT, VTT, TXT, HTML, PDF, JSON and DOCX. Queue a folder, leave it running overnight, and check the results in the morning.

For interview shows: noScribe #

If most episodes are conversations, noScribe is worth the extra setup. It was built for transcribing research and journalism interviews. It uses pyannote to separate speakers and has an editor for checking the transcript against the audio. It’s slower than Buzz on machines without a strong GPU.

Using Whisper from the command line on Windows #

If you’d rather script it:

1. Install Python and ffmpeg. OpenAI’s tool needs ffmpeg to read audio.
2. Run pip install -U openai-whisper.
3. Run whisper episode.mp3 --model turbo --language English.

Per OpenAI’s README, turbo needs about 6 GB of video memory. Use small or medium on a PC without a capable NVIDIA GPU. whisper.cpp is lighter on memory and fast on CPUs. It builds with CMake and its whisper-cli tool expects 16-bit WAV input, which ffmpeg can produce.

GPU or CPU: how long will it take? #

Speed depends on model size and hardware far more than on the app. Small models run comfortably on any recent CPU. Large models are much faster with a supported GPU. On a laptop without one, a long episode on a large model can take longer than the episode itself. If a transcript comes back too rough, step up a model size. If it’s too slow, step down and put more effort into the editing pass.

For ways to improve the audio before it reaches the model, see how to improve speech to text accuracy for podcasts.

Dictation tools for scripts and show notes on Windows #

Transcribing episodes is one job. Writing scripts and show notes by voice is another. For that, you want dictation that types into any app:

• Voice Access is built into Windows 11 (version 22H2 and later). Microsoft says it lets you control your PC and author text using only your voice and without an internet connection.
• Handy is free and open source. Hold a shortcut, speak, and Whisper or Parakeet transcribes on your PC and pastes the text where your cursor is.
• Dragon Professional v16 is a paid, Windows-only option for heavy dictation users who want deep voice commands.

Popular cloud dictation tools don’t belong on a privacy list. Wispr Flow, for example, says transcription always happens in the cloud. For a fuller comparison of dictation tools, see best offline speech recognition apps for Windows PC.

When your phone is the simpler route #

A lot of podcast audio starts on a phone: guest backup recordings, field interviews, voice memos with episode ideas. Moving those to a PC just to transcribe them is an extra step. Private Transcribe is an iPhone and Android app, not a Windows program. It runs Whisper on the phone, imports m4a, mp3, wav, aac, mp4 and mov files up to 90 minutes, and never uploads audio. With the one-time Pro purchase it exports .txt and .srt, so you can send just the finished text to your PC.

It doesn’t label speakers or batch-process, so for full multi-voice episodes a desktop app on Windows remains the better tool. For more on tool choices, see best offline speech to text for podcast creators.

Frequently asked questions #

Is Buzz safe to install on Windows? #

Buzz is open source under the MIT license, with its code public on GitHub. The installer isn’t code-signed, so Windows warns you when you run it. That’s a signing issue, not a sign of malware. Download it only from the links on the project’s own page.

Does Microsoft Word transcribe audio offline? #

No. Word’s Transcribe feature uploads recordings and stores them on OneDrive. Microsoft 365 subscribers get 300 minutes of uploaded audio a month. For offline transcription on Windows, use a local Whisper app instead.

What’s the best free offline podcast transcriber for Windows? #

Buzz for simplicity, Vibe for batches and export formats, and noScribe for interview shows where speaker labels matter. All three are free, open source and run Whisper on your PC.

Can Windows Voice Access transcribe an audio file? #

No. Voice Access handles live dictation and PC control from your microphone. To turn a recorded episode into text, you need a file transcription tool such as Buzz, Vibe or noScribe.