A macOS menu bar app that turns your voice into text. Hold a key, talk, release, and the transcription gets pasted wherever your cursor is. Works in any app.
Download SimpleDictation-1.3.0.dmg -- open the DMG and drag SimpleDictation to your Applications folder. No dependencies required.
Requires macOS 14.0 or later.
- You hold down the fn key (or option key)
- Talk into your mic
- Release the key
- The text gets typed out wherever your cursor is -- chat boxes, text editors, search bars, anywhere
Double-tap the key to press Enter (submits forms, sends messages).
There's also a floating mic button on screen you can click instead of using the hotkey.
- Go to Releases
- Download
SimpleDictation-1.3.0.dmg - Open the DMG and drag SimpleDictation to Applications
- Launch from Applications
Requires Homebrew, XcodeGen (brew install xcodegen), and Xcode Command Line Tools (xcode-select --install).
git clone https://github.com/cfranci/SimpleDictation.git
cd SimpleDictation
./build.shThe build script generates the Xcode project, compiles the app, and launches it automatically.
On first launch, macOS will ask for three permissions. Say yes to all of them:
- Microphone -- so it can hear you
- Accessibility -- so it can type text into other apps and handle the clipboard history
- Speech Recognition -- for the Apple Speech engine
If you accidentally denied one, go to System Settings > Privacy & Security and enable them for SimpleDictation.
The app lives in your menu bar (top right of your screen). Click the icon to switch engines, change your hotkey, or adjust settings. Right-click the floating mic button for the same menu.
The app comes with multiple speech-to-text engines. You can switch between them from the menu bar.
| Engine | Download Size | Speed | Accuracy | Notes |
|---|---|---|---|---|
| Apple Speech | Nothing to download | Instant | Good | Uses built-in macOS dictation |
| Whisper Tiny | ~40 MB | ~1s | Fair | Fastest local model |
| Whisper Base | ~140 MB | ~2s | Good | Good balance |
| Whisper Small | ~460 MB | ~4s | Better | Recommended for most people |
| Whisper Medium | ~1.5 GB | ~7s | Great | Best for long-form |
| Distil-Whisper Large v3 | ~594 MB | ~5s | Great | High accuracy, reasonable speed |
| Distil-Whisper Large v3 Turbo | ~600 MB | ~9s | Best | Most accurate overall |
| Whisper Large v3 Turbo | ~1.5 GB | ~5s | Best | Latest OpenAI model, streaming-optimized |
| Whisper Large v3 Turbo (632MB) | ~632 MB | ~5s | Best | Compressed version, best bang for buck |
| Moonshine Tiny | Bundled | ~1s | Fair | Comes with the app |
Start with Apple Speech -- it works instantly with no downloads. If you select a model that hasn't been downloaded yet, the app uses Apple Speech as a fallback while the model downloads in the background, then switches over automatically.
Models that aren't downloaded yet appear grayed out in the menu. Select one to start the download. You'll see a notification when it's ready.
| Action | How |
|---|---|
| Record | Hold fn key (or option, configurable) |
| Stop and paste | Release the key |
| Submit / press Enter | Double-tap the key |
| Toggle recording | Click the floating mic button |
| Submit via button | Double-click the floating mic button |
| Open settings | Right-click the floating mic button, or click the menu bar icon |
SimpleDictation also includes clipboard history. The last 10 things you copied are saved.
- Hold Cmd, then tap V repeatedly to cycle through your clipboard history
- Works in Chrome, Safari, Notes, TextEdit, and most apps
- Hold-to-talk dictation with automatic paste
- 10 speech engines from instant to high-accuracy
- Floating mic button with recording glow, audio level ring, and resizable icon (10-50px)
- Clipboard history cycling (last 10 copies, Cmd+V to cycle)
- Incremental mode to see partial transcription as you speak
- 16 languages including English, Spanish, French, Chinese, Japanese
- Double-tap to submit for chat boxes and forms
- Auto-fallback to Apple Speech while models download
- Model management with download progress notifications
If you pull updates:
cd SimpleDictation
git pull
./build.sh