VoicePaste

Hold a hotkey, speak, release — your words are transcribed and pasted instantly.

Push-to-talk speech-to-text for your desktop. Powered by Groq Whisper API.

Download • Features • How it works • Configuration • Development

---

Download

Go to the latest release and download:

File	Platform
VoicePaste-Setup-x.x.x.exe	Windows installer (recommended)
VoicePaste-x.x.x-portable.exe	Windows portable — no installation needed
VoicePaste-x.x.x.dmg	macOS (Intel & Apple Silicon)
VoicePaste-x.x.x.AppImage	Linux
VoicePaste-x.x.x.deb	Debian / Ubuntu

After installing, you'll need a Groq API key from console.groq.com/keys.

Features

• Push-to-talk — hold your hotkey, speak, release. Text is transcribed and pasted into any active application
• Blazing fast — Groq's Whisper Large v3 Turbo processes speech in under a second
• Multi-language — supports 50+ languages with automatic language detection and code-switching (mix languages naturally)
• Smart audio processing — built-in noise filtering, compression, and gain boost for clearer transcription
• Active window context (experimental) — detects your active app (VS Code, terminal, email) and hints the AI for better accuracy with domain-specific terms
• Audio ducking — automatically lowers system volume while you speak
• History — log of all transcriptions with full API metadata (processing time, detected language, audio duration)
• Clipboard management — optionally restores your previous clipboard content after pasting
• System tray — runs quietly in the background, always ready
• Auto-start — launches with your operating system
• 8 UI languages — English, Polish, German, Spanish, French, Chinese, Japanese, Portuguese
• Settings export/import — backup and restore your configuration as JSON
• Auto-updates — checks for new versions automatically via GitHub Releases
• Cross-platform — Windows, macOS, and Linux
• Privacy-first — your API key is encrypted locally. Audio is sent only to Groq's API, never stored remotely

How it works

Hold hotkey → Speak → Release hotkey → Text appears where your cursor is

1. Press and hold your configured hotkey (default: Ctrl + Win on Windows, Ctrl + Cmd on macOS)
2. Speak naturally — an overlay shows recording status
3. Release the hotkey — the overlay switches to "Transcribing..."
4. In under a second, the transcribed text is pasted at your cursor position

VoicePaste captures audio from your microphone, processes it through a real-time Web Audio API pipeline (high-pass filter, compressor, gain boost), sends it to Groq's Whisper API, and simulates paste to insert the result.

Configuration

On first launch, VoicePaste guides you through setup:

Setting	Description
Groq API Key	Required. Get one at console.groq.com/keys
Hotkey	Key combination to hold while speaking (default: Ctrl + Win / Ctrl + Cmd)
Microphone	Select your audio input device — with live level indicator
Primary language	Your main language for transcription
Spoken languages	All languages you use — enables auto-detection when multiple are selected
Audio enhancement	Toggle noise filtering and volume boost (on by default)
Keep clipboard	Whether to leave transcribed text in clipboard after pasting
Active window context	(Experimental) Send active window title to improve transcription accuracy
Launch at startup	Start VoicePaste with your OS
UI language	Interface language (8 available)

All settings can be exported/imported as JSON for backup or sharing across machines.

Tech stack

• Electron — cross-platform desktop framework
• Vue 3 + Naive UI — reactive UI with dark theme
• TypeScript — type-safe codebase
• Groq API — Whisper Large v3 Turbo for fast speech-to-text
• uiohook-napi — global hotkey detection
• better-sqlite3 — local history storage
• electron-vite — build tooling
• vue-i18n — internationalization
• Web Audio API — real-time audio processing pipeline

---

Development

Prerequisites

• Node.js >= 20
• Windows, macOS, or Linux

Setup

git clone https://github.com/ksroga/voice-paste.git
cd voice-paste
npm install

Development mode

# Run the app in dev mode (hot reload)
npm run rebuild:electron
npm run dev

# Linting, type checking, unit tests (no GUI needed)
npm run lint
npm run typecheck
npm run test:unit

Note: Native modules (uiohook-napi, better-sqlite3) compile per platform. Run npm run rebuild:electron after npm install to recompile for Electron's Node.js version.

Project structure

src/
├── main/                  # Electron main process
│   ├── index.ts           # App lifecycle, IPC handlers, tray, hotkey flow
│   ├── hotkey.ts          # Global hotkey detection (uiohook-napi)
│   ├── groq.ts            # Groq Whisper API client
│   ├── paste.ts           # Clipboard + paste simulation (Windows/macOS/Linux)
│   ├── db.ts              # SQLite history storage
│   ├── config.ts          # Settings (JSON file + safeStorage encryption)
│   ├── overlay.ts         # Recording/transcribing overlay window
│   ├── sounds.ts          # Start/stop sound effects (cross-platform)
│   ├── audio-control.ts   # System volume ducking
│   ├── active-window.ts   # Active window detection (Windows/macOS/Linux)
│   └── updater.ts         # Auto-update via GitHub Releases
├── preload/
│   └── index.ts           # IPC bridge (contextBridge)
└── renderer/
    ├── src/
    │   ├── App.vue        # Root component + audio recording pipeline
    │   ├── components/    # HistoryView, SettingsView, OnboardingView, HotkeyRecorder
    │   ├── i18n/          # Translations (8 languages)
    │   └── languages.ts   # Whisper language options
    └── overlay.html       # Recording overlay (standalone)

Scripts

Command	Description
npm run dev	Start in development mode (hot reload)
npm run build	Build for production
npm run package	Package for Windows (installer + portable)
npm run package:mac	Package for macOS (dmg)
npm run package:linux	Package for Linux (AppImage + deb)
npm run package:all	Package for all platforms
npm run lint	Run ESLint
npm run typecheck	Run TypeScript type checking
npm run test:unit	Run unit tests (Vitest)

Building

# Build + package for your current platform
npm run build
npm run package          # Windows
npm run package:mac      # macOS (must run on macOS)
npm run package:linux    # Linux

Output goes to dist/.

Releasing

Releases are automated via GitHub Actions. To publish a new version:

# 1. Update version in package.json
# 2. Update CHANGELOG.md
# 3. Commit and tag
git add -A
git commit -m "Release v0.1.0"
git tag v0.1.0
git push origin main --tags

GitHub Actions will build on Windows, macOS, and Linux, then publish a GitHub Release with all binaries.

---

Contributing

Contributions are welcome! Please see CONTRIBUTING.md for guidelines.

1. Fork the repository
2. Create a feature branch (git checkout -b feature/my-feature)
3. Commit your changes
4. Push to the branch (git push origin feature/my-feature)
5. Open a Pull Request

For bugs or feature requests, please open an issue.

License

MIT © 2026 Konrad Sroga