4584d30e02
Quantized models: add Small q5_1, Medium q5_0, Large v3 Turbo q5_0/q8_0 and Large v3 q5_0 to the catalog. Turbo q5_0 keeps whisper-server at ~0.7 GB instead of ~1.75 GB with the same speed. Installed models show their catalog name; the first-run wizard uses a popup instead of one button per model. Download feedback: a progress window with bar, size, speed, time left, cancel and retry, followed by "loading" and "ready" steps. Progress also shows in the menu bar icon, the status menu and the Settings button, and the Settings model list refreshes when done. Only one download runs at a time, an existing model is replaced only after a successful download, and HTTP errors are reported instead of saving the error page as a model. Model warm-up: start loading the model when recording starts instead of after it stops. After 2+ minutes idle a tiny silent request pages the model back in while the user talks; a dead server is restarted, and a server that is still loading is awaited rather than loading the model a second time via whisper-cli. If loading takes over 2.5 s the recording HUD explains why this transcription is slower. Server start no longer overwrites the recording/transcribing menu bar icon. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
WhisperDictate
A simple menu bar app for voice dictation using OpenAI Whisper (local, offline).
Platforms
| Platform | Language | Status |
|---|---|---|
| macOS | Swift | ✅ Ready |
| Linux | Rust | 🔜 Planned |
| Windows | C# | 🔜 Planned |
macOS
What's new in 1.3.0
- ⚡ Warm transcription server — the Whisper model stays loaded in a background
whisper-server, so each dictation is a fast local call (~0.5s) instead of reloading the 1.5–1.6 GB model on every utterance - 📊 Live mic-level HUD while recording, which warns ("No audio input?") if the microphone stays silent
- 📖 Custom dictionary — an editable Heard → Replacement table that fixes names/jargon in the transcript and also biases recognition
Features
- 🎤 Global hotkey (⌃⌥D) to start/stop recording
- ⚡ Fast transcription via a warm whisper-server (model stays loaded — no per-dictation reload), with automatic fallback to
whisper-cli - 📊 Live mic-level meter while recording, with a "no audio input" warning
- 📖 Custom dictionary (heard → replacement) that also biases recognition
- 🔒 Fully offline - uses local Whisper model
- ⌨️ Automatic paste into any focused app
- 📋 Clipboard preservation - your copied content is restored after paste
- ⚙️ Settings window with model selection dropdown
- 📥 Built-in model downloader with progress indicator
- 🚀 Launch at login support
- 🔊 Sound feedback (optional)
- 📦 Self-contained - whisper-cli and whisper-server bundled in the app
Requirements
- macOS 13.0+
- Apple Silicon (M1/M2/M3) or Intel Mac
Quick Install (Download)
- Download the latest DMG from Releases
- Open the DMG and drag WhisperDictate to Applications
- Launch WhisperDictate
- On first run, select and download a Whisper model
- Grant permissions (Microphone + Accessibility)
Build from Source
# Clone the repository
git clone https://github.com/hariel1985/WhisperDictate.git
cd WhisperDictate/macos
# Install whisper-cpp (required for bundling)
brew install whisper-cpp
# Build and install to /Applications
make install
Build Commands
| Command | Description |
|---|---|
make build |
Compile the app and bundle whisper-cli |
make install |
Build and install to /Applications |
make run |
Build and run |
make dmg |
Create distributable DMG |
make clean |
Remove build artifacts |
Usage
- Launch WhisperDictate from Applications
- Look for the 🎤 icon in your menu bar
- Press ⌃⌥D (Control + Option + D) to start recording
- Speak (icon changes to 🔴)
- Press ⌃⌥D again to stop and transcribe
- Text is automatically pasted where your cursor is
Settings
Click the menu bar icon → Settings to configure:
- Language: Auto-detect or 31 supported languages
- Model: Select from installed models or download new ones
- Dictionary: Add Heard → Replacement pairs to fix names/jargon (use
\nfor a newline); the replacement terms also improve recognition - Sound feedback: Toggle audio feedback on/off
- Launch at login: Start automatically when you log in
Whisper Models
Download models directly from the app or manually:
| Model | Size | Speed | Accuracy | Best For |
|---|---|---|---|---|
| Tiny | 75 MB | ~1 sec | Basic | Quick tests, simple phrases |
| Base | 142 MB | ~2 sec | Good | Clear speech, quiet environment |
| Small | 466 MB | ~3 sec | Better | General use, some accents |
| Medium | 1.5 GB | ~5 sec | Great | Accents, noisy audio |
| Large v3 Turbo | 1.6 GB | ~4 sec | Best | Recommended - fast like Medium, accurate like Large |
| Large v3 | 3.1 GB | ~8 sec | Maximum | Difficult audio, max accuracy |
Models are stored in ~/.whisper-models/
With the warm server (1.3.0+), the selected model loads once at startup (a few seconds); after that each dictation skips the reload, so the speeds above are roughly the per-dictation transcription time.
Audio Feedback
- 🔔 Tink - Recording started
- 🔔 Pop - Recording stopped, processing
- 🔔 Glass - Success, text pasted
- 🔔 Basso - Error
Permissions
Grant these in System Settings → Privacy & Security:
| Permission | Why it's needed |
|---|---|
| Microphone | To record your voice for transcription |
| Accessibility | To simulate ⌘V keystroke and paste text into any app. macOS requires this permission for apps that send keyboard events to other applications. |
Note
: After reinstalling or updating, you may need to remove and re-add the app in Accessibility settings.
Security
- All processing is done locally - no data leaves your device
- The bundled
whisper-serveris bound to127.0.0.1(loopback only) and is never exposed on the network - Audio files are stored in private temp directory and deleted after transcription
- Clipboard is cleared after paste (transcript doesn't remain accessible)
- Original clipboard content is preserved and restored after paste
- Input validation prevents command injection
- No network access except for optional model downloads from Hugging Face
License
MIT License