Free, Local and Open Source speech-to-text application powered by Whisper. Your voice stays on your machine. Compatible with Windows, Linux and MacOS
- 100% Free - For ever
- 100% Local - No internet required, no data leaves your computer
- Floating Overlay - Minimal, always-on-top microphone button
- Auto-copy - Text is inserted directly in your clipboard
- Global Shortcut - Cmd/Ctrl+Shift+Space from anywhere
- Multi-language - 12+ languages with auto-detection
Please check the Release
| OS | Requirements |
|---|---|
| macOs | Must have GPU Metal optimization installed by default since 2012, fully compatible since 2021 with the M series |
| Windows | Must have Vulkan installed by default with NVidia, AMD and Intel driver since 2012 |
| Linux debian and ubuntu | sudo apt install mesa-vulkan-drivers nvidia-driver-550 |
| Linux fedora | sudo dnf install mesa-vulkan-drivers akmod-nvidia |
- Click the microphone button to start recording
- Speak
- Click again to stop and transcribe
- Text is automatically inserted at your cursor
Shortcut: Cmd+Shift+Space (macOS) or Ctrl+Shift+Space (Windows/Linux)
# Install dependencies
npm install
# Whisper models for local dev (recommended — saves you from the
# in-app auto-download every time you wipe the build).
# Models are downloaded by released apps from the models-v1 GitHub
# Release; in dev mode, the app reads them from src-tauri/models/
# directly.
mkdir -p src-tauri/models
curl -L -o src-tauri/models/ggml-small.bin \
https://github.com/AccessDevops/S2Tui/releases/download/models-v1/ggml-small.bin
curl -L -o src-tauri/models/ggml-large-v3-turbo.bin \
https://github.com/AccessDevops/S2Tui/releases/download/models-v1/ggml-large-v3-turbo.bin
# Run
npm run tauri dev
# Build (no model bundling; produced binary downloads them on first launch)
npm run tauri build- Frontend: Vue 3, TypeScript, Tailwind CSS
- Backend: Tauri 2, Rust
- Speech: whisper-rs (Whisper.cpp)
MIT