A desktop transcription app that runs OpenAI Whisper models locally on your machine. No cloud, no API keys, no subscriptions.
- Fully local — your audio never leaves your computer
- CUDA acceleration — fast transcription on NVIDIA GPUs
- CPU fallback — works on any Windows PC
- Multiple output formats — TXT, SRT, VTT, JSON
- 99 languages supported
- Model selection — tiny to large, balance speed vs accuracy
Get the latest installer from Releases.
The lite installer (~77MB) downloads the required dependencies during setup:
- Python runtime
- PyTorch (with CUDA 12.4 support)
- OpenAI Whisper
- FFmpeg
Internet connection required during installation. Total download during install: ~2.5GB.
Note: Windows may block the installer because it's unsigned. Right-click the exe → Properties → Unblock → OK, then run it.
- Open Transcriptor
- Select an audio or video file
- Choose an output folder
- Pick a model (medium is a good default)
- Select output formats
- Click Start Transcription
The first time you use a model, it will be downloaded (~1.5GB for medium). After that, transcriptions start immediately.
MP3, MP4, WAV, FLAC, OGG, M4A, WEBM, MKV, AVI, MOV
| Format | Description |
|---|---|
| TXT | Plain text transcript |
| SRT | SubRip subtitles |
| VTT | WebVTT subtitles |
| JSON | Full Whisper result object |
| Model | Size | Speed | Quality |
|---|---|---|---|
| tiny | 75 MB | Fastest | Basic |
| base | 142 MB | Fast | Fair |
| small | 462 MB | Moderate | Good |
| medium | 1.5 GB | Slow | Great |
| large | 2.9 GB | Slowest | Best |
| turbo | 1.6 GB | Fast | Great |
English-only variants available: tiny.en, base.en, small.en, medium.en
Models are downloaded on first use to %APPDATA%/Transcriptor/models/.
- Windows 10/11 (64-bit)
- 8 GB RAM minimum (16 GB recommended for large models)
- Internet connection for initial setup and model downloads
- NVIDIA GPU with CUDA support
- Latest NVIDIA drivers
AMD GPUs are not supported.
- Node.js 18+
- Python 3.10+
- FFmpeg on PATH
npm install
cd python
pip install -r requirements.txtnpm run devscripts\build-lite.batElectron (TypeScript/React)
└── spawns → Python worker
└── uses → openai-whisper + PyTorch + FFmpeg
The Electron app handles the UI and process management. Transcription is performed by a Python worker using the official OpenAI Whisper package.
Copyright (c) 2026 Jonathan Leibbrandt
This application is licensed under the Apache License, Version 2.0. See the LICENSE file for details.
This project uses OpenAI Whisper, licensed under the MIT License.