The four free tools
1. Buzz
Open-source desktop app. Installer, a window, drag in a file, get text. No speaker labels, basic export. The right answer for most people who want free.
2. WhisperX
Command line. Adds speaker labels and word timestamps to Whisper. Needs Python, a Hugging Face token and, for GPU speed, matching CUDA. An afternoon of setup.
3. Whisper CLI
OpenAI's reference tool. Command line, Python, no speaker labels. The cleanest option if you script your workflow.
4. whisper.cpp
A fast C++ port that runs well on CPUs. Command line, some prebuilt Windows binaries. No speaker labels. Good on older PCs without a GPU.
Side by side, including the paid option
| Buzz | WhisperX | Whisper CLI | whisper.cpp | Scrieb | |
|---|---|---|---|---|---|
| Price | Free | Free | Free | Free | Paid, trial |
| Interface | App | Terminal | Terminal | Terminal | App |
| Speaker labels | ✗ | ✓ | ✗ | ✗ | ✓ |
| DOCX / PDF export | ✗ | ✗ | ✗ | ✗ | ✓ |
| Local summary | ✗ | ✗ | ✗ | ✗ | ✓ |
| GPU without manual setup | ✓ | ✗ | ✗ | ✗ | ✓ |
| Video files | ✓ | via ffmpeg | via ffmpeg | via ffmpeg | ✓ |
| Also on Mac | ✓ | ✓ | ✓ | ✓ | ✓ |
Which one to pick
- Single-voice recordings, lectures, dictation, memos: Buzz. Free and done.
- Interviews or meetings, and you know Python: WhisperX. Free speaker labels, some setup.
- Old PC without a graphics card: whisper.cpp, or Buzz with a small model.
- Interviews or meetings, no terminal, need a finished document: that is the gap Scrieb fills, and the only case where I'd tell you to pay.
What all of them need from your PC
Windows 10 or 11, 8 GB of RAM, and patience if there is no dedicated GPU: CPU-only transcription runs at roughly real time. An NVIDIA RTX 20-series or newer takes a one-hour file down to 4–12 minutes. Details in offline transcription for Windows.
Frequently asked questions
- What is the best free offline transcription software for Windows?
- Buzz, if you want an app with a window: free, open source, drag in a file, get text. WhisperX, if you can use a terminal and need speaker labels. Whisper's own command line if you script things. All three run OpenAI's Whisper model on your PC with no upload.
- Is there a free offline audio-to-text converter for Windows with speaker labels?
- WhisperX is the only free option that labels speakers, and it is a command-line tool that needs Python, CUDA drivers for GPU speed and a Hugging Face account for the speaker model. Free desktop apps like Buzz do not label speakers. Scrieb does, but it is paid after the trial.
- Do free tools need a graphics card?
- No, they run on the CPU, slowly: roughly real time or slower. An NVIDIA RTX card makes every one of them 5 to 15 times faster, but getting GPU acceleration working in the command-line tools means installing the right CUDA build yourself.
- Is Windows' built-in dictation an option?
- Windows voice typing transcribes live speech from a microphone, not audio files, and on most systems it sends audio to Microsoft. It is not a tool for transcribing recordings.
- Why would I pay for Scrieb if free tools exist?
- For speaker labels without a terminal, DOCX and PDF export, a summary generated locally, GPU acceleration that works out of the box, and one app for Windows and Mac. If none of that matters to you, use Buzz.
- Is Scrieb free?
- No. There is a free trial, then a subscription. This page exists because people searching for free tools deserve a straight answer, and the free tools are good.