What's Voice Type?

A system-wide Speech-to-Text for Linux. Press a key, speak, done — text lands wherever you're focused. Totally free, with unmatched speed and accuracy.

How it works: Chrome or Chromium runs quietly in the background using its built-in Web Speech API. No local models, no paid services, no startup lag.

Features

• Instant Dictation: Press a hotkey and start speaking. Text appears in real-time in whatever window you have focused in the system.
• Auto-Correction: If the AI corrects a word mid-sentence, it automatically backspaces and types the corrected version.
• System-Wide: Works in any application - text editors, browsers, terminals, messaging apps, etc.
• Visual & Audio Feedback: Get instant confirmation when listening starts, stops, or if there's an error.
• Zero Latency: Browser runs persistently in the background - no startup delay when you press the hotkey.
• Totally free: Uses Chrome's built-in Web Speech API - so no paid services required.
• No local models required: The browser Web Speech API handles the transcription process entirely, so there's no need to install and setup models locally (like vosk or whisper).

Installation

Prerequisites

Before installing Voice Type, make sure you have:

• Linux - any distribution
• A desktop environment - GNOME preferred, but any DE works
• Google Chrome or Chromium - required for the Web Speech API (Chrome is used by default)
• A working microphone

Via Flatpak (Universal Compatibility)

Voice Type is available as a Flatpak for maximum compatibility across all Linux distributions and Desktop Environments. Due to the self-contained and isolated nature of flatpaks, it is the most compatible and easiest installation method, so no need to manually install system dependencies.

Flatpak Limitations: The Flatpak version supports Chrome and Chromium via the --browser flag, but does not support custom browser paths via --browser_path. If you need to use a custom browser path (e.g., Chrome Beta, Dev, or a non-standard installation), use the binary or npm installation instead.

git clone https://github.com/eriknovikov/voice-type.git
cd voice-type
chmod +x ./flatpak/build.sh && ./flatpak/build.sh

Binary Installation

The script takes care of downloading and setting up everything for you, and will let you know which system dependencies are missing.

curl -fsSL https://github.com/eriknovikov/voice-type/releases/latest/download/install.sh | bash

Via npm

Voice Type is available as an npm package for easy installation on any Linux system with Node.js installed.

npm install --global voice-type-cli@latest

Required System Dependencies

Whether you use the binary directly or via npm, VoiceType requires google-chrome-stable (or chromium), dotool, and paplay installed in the system. Their installations depend on your distro, so install them manually.

Note about dotool: On some distributions, dotool is available in community or third-party repositories (COPR for Fedora, AUR for Arch, etc.). If not available in your distro's repos (like Ubuntu-Debian), you'll need to build it from source. See dotool source for instructions. If this sounds like too much of a hassle, just use the flatpak.

Uninstalling

Flatpak

To remove Voice Type installed via Flatpak:

flatpak uninstall org.github.eriknovikov.VoiceType

This will remove the application and all its associated data. If you also want to remove any leftover configuration files:

flatpak uninstall --unused

npm

To remove Voice Type installed via npm:

npm uninstall --global voice-type-cli

This will remove the globally installed package from your system.

Binary

To remove Voice Type installed via the binary installation script:

sudo rm -rf /usr/local/share/voice-type /usr/local/bin/voice-type

This removes the application files and the executable symlink. If you also want to remove any user-specific configuration or cache files:

rm -rf ~/.config/voice-type ~/.cache/voice-type

Note: After uninstalling, you may also want to remove any keyboard shortcuts you set up for Voice Type from your desktop environment's keyboard settings.

Troubleshooting

Problem: dotool: command not found

• Solution: Install dotool using your package manager (see above)

Problem: voice-type: command not found

• Solution: Make sure /usr/local/bin is in your PATH. Add this to your ~/.bashrc or ~/.zshrc:

  export PATH="$PATH:/usr/local/bin"

Problem: Microphone not detected

• Solution: Check your system audio settings and ensure your microphone is properly connected

Problem: Chrome not found

• Solution: Install Google Chrome or Chromium. Voice Type uses Chrome by default, but you can specify Chromium with --browser chromium. For custom browser paths, use the binary or npm installation (Flatpak does not support --browser_path).

Usage

Set up Keyboard Shortcuts

Voice Type uses HTTP endpoints to control dictation. You'll need to bind these to keyboard shortcuts for easy access. We suggest F9 and F10, but you can use whatever you want.

| Key | Action | Command | | ---------- | -------------------------- | ---------------------------------- | | F9 | Start dictation [required] | curl http://127.0.0.1:3232/start | | F10 | Stop dictation [required] | curl http://127.0.0.1:3232/stop | | Ctrl + F9 | Start daemon [optional] | depends on installation | | Ctrl + F10 | Stop Daemon [optional] | curl http://127.0.0.1:3232/exit |

If you use GNOME, go to Settings -> Keyboard -> View and Customize Shortucts -> Custom Shortcuts. Else, check manually how to set up keyboard shortcuts in your Desktop Environment.

Then, run the daemon, depending on how you installed it.

#flatpak
flatpak run org.github.eriknovikov.VoiceType

#npm or binary
voice-type

To see all the different options, pass --help or -h to the command. You can customize the language, browser type (Chrome or Chromium), custom browser path, whether to disable sounds and text notifications for when the daemon or mic starts / stops, and whether to run voice-type in dettached mode.

After starting the daemon, move your cursor to any textbox within your system where you want to VoiceType into. then hit F9, speak whatever you wish, and text results will be inserted in real time. to stop listening, press F10.

You might also find useful to have keyboard shortcuts for starting or stopping the daemon, but these are optional and totally up to you.

Contributing

Voice Type is totally free and open source, and you can contribute in many ways! Feel free to report/fix a bug or issue, improve documentation, and suggest or add a new feature.