Whisper Studio is an independent desktop interface for WhisperX and faster-whisper.
complete with apps for macOS, Windows, Linux
A cross-platform desktop application for AI-powered audio and video transcription using [WhisperX](https://github.com/m-bain/whisperX) and Whisper models. Built with Electron, React, TypeScript, and Vite.
Whisper Studio runs entirely on your machine — no cloud services, no data leaves your device. It manages a local Python environment, checks required tools, installs missing packages when possible, transcribes audio or video files, and exports to SRT, VTT, TXT, TSV, or JSON.
- Transcribe audio and video files locally using WhisperX
- GPU (CUDA) and CPU compute modes
- Real-time transcription progress and live log output
- Interactive transcript editor with search and replace
- Export to SRT, VTT, TXT, TSV, and JSON
- Model manager — download and delete Whisper models in-app
- Prerequisite checker and installer for Python, FFmpeg, CUDA, WhisperX, and PyTorch
- Dark / light theme
npm install
npm run devElectron Builder is configured for:
- macOS:
dmg,zip - Windows:
nsis,portable - Linux:
AppImage,deb
Code signing, notarization, update feeds, and platform-specific icons should be added before publishing production releases.