Speed of Sound - Ubuntu Voice Dictation Setup

speed-of-sound-setup.md

Speed of Sound - Ubuntu Voice Dictation Setup

Local Whisper-based voice typing for Linux, similar to macOS dictation.

Installation

# Install Flatpak & Speed of Sound
sudo apt update && sudo apt install -y flatpak
flatpak remote-add --if-not-exists flathub https://flathub.org/repo/flathub.flatpakrepo
flatpak install -y flathub io.speedofsound.SpeedOfSound

Auto-start at Login

mkdir -p ~/.config/autostart
cat > ~/.config/autostart/speed-of-sound.desktop << 'EOF'
[Desktop Entry]
Type=Application
Name=Speed of Sound
Comment=Voice typing for the Linux desktop
Exec=flatpak run io.speedofsound.SpeedOfSound
Icon=io.speedofsound.SpeedOfSound
Terminal=false
Categories=Utility;
X-GNOME-Autostart-enabled=true
EOF

Global Keyboard Shortcut (Super+V)

Create trigger script:

mkdir -p ~/speedofsound
cat > ~/speedofsound/trigger.sh << 'EOF'
#!/bin/bash
gdbus call --session \
  --dest io.speedofsound.SpeedOfSound \
  --object-path /io/speedofsound/SpeedOfSound \
  --method org.gtk.Actions.Activate "trigger" [] {}
EOF
chmod +x ~/speedofsound/trigger.sh

Then go to Settings → Keyboard → View and Customize Shortcuts → Custom Shortcuts → +

• Name: Speed of Sound
• Command: /home/<your-username>/speedofsound/trigger.sh
• Shortcut: press Super+V

Usage

1. Launch Speed of Sound (flatpak run io.speedofsound.SpeedOfSound)
2. Minimize to system tray
3. Press Super+V in any app, speak, press Super+V again
4. Text appears at cursor

Model Selection

In Speed of Sound Preferences → ASR Model:

• tiny.en — fastest, good for CPU
• base.en — balanced speed/accuracy
• small.en — more accurate, needs GPU