中文版 | English
System-wide voice-to-text for macOS. Press a hotkey to record, release to transcribe with OpenAI Whisper and instantly paste into any app.
- Press
Ctrl+Option+V— recording starts, floating overlay appears - Press again (or click STOP) — audio is sent to OpenAI Whisper
- Transcribed text is pasted into your currently focused app
- Native Swift app — no Python, no Electron, no dependencies
- Menu bar + Dock — idle, recording, transcribing, paraphrasing states
- Global hotkey — customizable via Settings (default: Ctrl+Option+V), registered via Carbon API — no accessibility permission needed for the hotkey itself
- Floating overlay — red pill-shaped recording indicator with pulsing dot and STOP button
- Auto-paste — copies to clipboard, optionally auto-pastes via ⌘V (requires Accessibility permission)
- AI paraphrase — rewrite transcription into natural Slack/work chat style before pasting
- Translation — translate output to 15+ languages (English, 繁體中文, 日本語, etc.)
- Transcript history — browse, copy, and delete past transcriptions
- Settings window — API key, model, language, temperature, hotkey, paraphrase config
- Multi-language — Cantonese, English, Mandarin, Japanese, and 97+ other languages
- macOS 13.0 (Ventura) or later
- Swift 5.9+ (included with Xcode 15+)
- OpenAI API key
# Clone
git clone https://github.com/agents-io/vibemic-native-macos.git
cd vibemic-native-macos
# Build
swift build -c release
# Create .app bundle
APP_DIR="VibeMic.app/Contents"
mkdir -p "$APP_DIR/MacOS" "$APP_DIR/Resources"
cp .build/release/VibeMic "$APP_DIR/MacOS/VibeMic"
cp VibeMic/Resources/Info.plist "$APP_DIR/Info.plist"
cp VibeMic/Resources/VibeMic.entitlements "$APP_DIR/Resources/"
# Codesign (required for microphone access)
codesign --force --deep --sign - \
--entitlements VibeMic/Resources/VibeMic.entitlements \
VibeMic.app
# Run
open VibeMic.appClick the menu bar icon to access Settings:
| Setting | Description |
|---|---|
| API Key | Your OpenAI sk-... key |
| Model | whisper-1, gpt-4o-transcribe, gpt-4o-mini-transcribe |
| Language | Auto-detect or specify (en, zh, ja, ko, etc.) |
| Prompt | Hint text for Whisper (e.g. expected languages) |
| Temperature | 0 (deterministic) to 1 (creative) |
| Shortcut | Click to capture a new global hotkey |
| Paraphrase | Toggle AI rewriting with customizable system prompt |
| Translate to | Translate output to a target language |
Settings and history are saved as config.json and history.json next to the .app bundle.
Alternatively, create a .env file next to the app:
OPENAI_API_KEY=sk-your-key-here
By default, VibeMic copies transcribed text to your clipboard. To enable automatic pasting:
- Open System Settings → Privacy & Security → Accessibility
- Add VibeMic.app and toggle it ON
- Restart VibeMic
- VibeMic Native Ubuntu — Ubuntu/Linux version
MIT
全系統語音轉文字 macOS 應用。按快捷鍵錄音,鬆開後自動用 OpenAI Whisper 轉錄,並即時貼上到任何應用程式。
- 按
Ctrl+Option+V— 開始錄音,浮動提示出現 - 再按一次(或點擊 STOP)— 音訊傳送至 OpenAI Whisper
- 轉錄文字自動貼上到你正在使用的應用
- 原生 Swift 應用 — 無 Python、無 Electron、無依賴
- Menu Bar + Dock — 閒置、錄音中、轉錄中、改寫中狀態
- 全域快捷鍵 — 可在設定中自訂(預設:Ctrl+Option+V)
- AI 改寫 — 將轉錄內容改寫成自然的 Slack/工作聊天風格
- 翻譯 — 輸出可翻譯為 15+ 種語言
- 轉錄記錄 — 瀏覽、複製、刪除過往的轉錄
- 多語言支援 — 廣東話、英文、普通話、日文等 97+ 種語言
git clone https://github.com/agents-io/vibemic-native-macos.git
cd vibemic-native-macos
swift build -c release詳細建置步驟請參考上方英文版。

