VoiType

Voice typing for Linux — press a key, speak, and text appears at your cursor.
Powered by Groq Whisper for lightning-fast speech recognition.

---

Why VoiType?

Linux has lacked a proper voice typing tool like Typeless on macOS. VoiType fills that gap — it's free, open-source, and works with any app on X11 and Wayland.

Hold a key. Speak. Release. Done.

Your speech is transcribed by Groq's Whisper API (blazing fast, free tier available), cleaned up by an LLM (filler removal, punctuation, grammar), and pasted at your cursor.

Features

• Voice Dictation — Hold Right Alt, speak, release. Formatted text appears at your cursor.
• Smart Rewrite — Select text, hold Left Alt + Right Alt, speak rewrite instructions.
• AI Formatting — Automatic filler removal ("um", "uh"), grammar, punctuation via LLM. Toggle off for raw transcription.
• Result Popup — Transcribed text shown in a floating popup. Click to copy if auto-paste doesn't work in certain apps.
• Microphone Selection — Choose your input device from Settings. Auto sample rate fallback.
• Visual Overlay — Recording (red pulse), Processing (blue spinner), Done (green check).
• Multi-language — Japanese, English, and any Whisper-supported language.
• X11 + Wayland — Auto-detects your session. Works with Sway, Hyprland, niri, GNOME, KDE, etc.
• System Tray — Quick access to settings, toggles, and hotkey reference.
• Privacy — No data stored. Audio is processed by Groq and discarded.

Quick Start

# 1. Clone and install
git clone https://github.com/wat-hiroaki/voitype.git
cd voitype
chmod +x setup.sh
./setup.sh

# 2. Log out and back in (required for input group)

# 3. Run
GROQ_API_KEY=your_key_here .venv/bin/voitype

Get your free API key at console.groq.com/keys.

Hotkeys

Key	Action
Right Alt (hold)	Voice dictation
Left Alt + Right Alt (hold)	Smart rewrite (select text first)
Esc	Cancel recording

Configurable via Settings (system tray > Settings).

How It Works

Hold key → Record → Groq Whisper API → LLM Format → Paste at cursor
              |                                          |
          microphone                              result popup

1. Record — Holding the hotkey captures audio from your microphone
2. Transcribe — Audio is sent to Groq's Whisper API (~0.5s latency)
3. Format — LLM cleans up filler words, self-corrections, adds punctuation
4. Paste — Formatted text is pasted at your cursor + shown in result popup

Configuration

Settings are stored in ~/.config/voitype/settings.json:

Setting	Description	Default
formatting_enabled	AI text formatting	true
sound_enabled	Sound feedback	true
hotkey_dictation	Dictation key	KEY_RIGHTALT
hotkey_modifier	Rewrite modifier	KEY_LEFTALT
audio_device	Microphone index (-1 = default)	-1
api_key	Groq API key	""

Requirements

• Linux (Debian/Ubuntu, Arch, Fedora)
• Python 3.10+
• Groq API key (free tier available)
• input group membership (setup.sh handles this)

Troubleshooting

Hotkeys not working

sudo usermod -aG input $USER
# Log out and log back in

API key dialog on every launch

echo 'export GROQ_API_KEY=your_key_here' >> ~/.bashrc
source ~/.bashrc

Or set it via the Settings dialog (saved permanently).

Right Alt conflicts with AltGr

Open Settings from the tray icon and change the dictation key to KEY_F13, KEY_SCROLLLOCK, or KEY_PAUSE.

No system tray icon

# Debian/Ubuntu
sudo apt install gir1.2-ayatanaappindicator3-0.1

# Arch
sudo pacman -S libayatana-appindicator

Audio not recording

arecord -d 3 test.wav && aplay test.wav

If that works, try selecting a different microphone in Settings.

Contributing

Contributions are welcome! Feel free to open issues or submit PRs.

License

MIT