Voice Type is a desktop push-to-talk speech-to-text app built with Rust, Tauri, React, and Bun.
Hold your hotkey, speak, release, and the app transcribes your speech with Groq Whisper and types the result into the active window.
- Version:
0.3.2 - Desktop shell:
Tauri 2 - Frontend:
React + TypeScript + Vite - Backend engine:
Rust + Tokio - Package manager:
Bun
- Push-to-talk transcription
- Global hotkey capture
- Floating recording HUD
- Appearance customization
- Optional grammar cleanup
- Text cleanup filters
- Local config storage
.
|- Cargo.toml
|- src/ # Rust backend engine
|- frontend/ # React desktop UI
|- src-tauri/ # Tauri shell and command bridge
|- package.json
`- assets/
- Rust
1.85+ - Bun
- Windows: WebView2
- Linux: Tauri/WebKitGTK prerequisites
- macOS: Xcode command line tools
bun install
bun run tauri:devUseful checks:
bun run build
cargo check --lib
cargo check --manifest-path src-tauri/Cargo.tomlConfig is stored at:
- Windows:
%APPDATA%\voice-type\config.json - macOS:
~/Library/Application Support/voice-type/config.json - Linux:
~/.config/voice-type/config.json
You can either:
- put the API key in the app settings, or
- leave
api_keyempty and useGROQ_API_KEY
- This repo is the current codebase.
- The older Python-based implementation should be treated as legacy.
- Bundled public image sources are documented in ASSET_SOURCES.md.
bun run tauri buildOutput: src-tauri/target/release/bundle/nsis/Voice Type_0.3.2_x64-setup.exe
Requires Windows SDK with makeappx.exe (already installed).
bun run tauri build
bun run msixOutput: msix-output/Boring877.VoiceType_0.3.2.0_x64__kg07y93afj4jj.msix
Upload the .msix file to Microsoft Partner Center to update the Store listing.
A symlink to the MSIX output folder is on the Desktop (
voice-type-msix) for quick access.