Entune guide · Linux

Dictation on Linux, on X11 and Wayland.

Hold a shortcut in any app, speak, and Entune pastes the text where you are. It runs as its own app on Linux, with local speech models or a cloud model on your own key.

Free and open source · X11 and Wayland · Local Whisper.cpp models

Updated

Install Entune See what it does →

Install it

Install Entune with the command below. On Linux it adds Entune to your applications menu and opens it.

uv tool install entune && entune

Needs uv, which installs Python for you, or pip with Python 3.12+. Free and open source.

Puts Entune in your applications menu and opens it. Linux needs a few system libraries; see the Linux guide.

Entune’s window and microphone use a few system libraries that Python packages cannot bring, and in a Wayland session it also uses wl-copy. If one is missing, entune names it and prints the command that installs it. On a fresh Ubuntu 24.04 desktop in a Wayland session, that is:

sudo apt install libminizip1t64 libportaudio2 libsnappy1v5 libxcb-cursor0 wl-clipboard

On Arch and other rolling-release distributions, install with uv’s own Python, so a system Python upgrade cannot break Entune:

uv tool install --managed-python entune && entune

Give it keyboard access

Get started asks for keyboard access, Linux’s counterpart of the Mac’s Input Monitoring and Accessibility. Entune reads your shortcut from the keyboard devices in /dev/input and types through a virtual keyboard on /dev/uinput, which works the same on X11 and on Wayland. Run this once in a terminal; Copy command in Entune copies it for you:

echo 'KERNEL=="uinput", GROUP="input", MODE="0660", OPTIONS+="static_node=uinput"' | sudo tee /etc/udev/rules.d/70-entune-uinput.rules && sudo modprobe uinput && sudo udevadm control --reload-rules && sudo udevadm trigger --sysname-match=uinput && sudo usermod -aG input "$USER"

It adds you to the input group and lets that group use /dev/uinput. Log out and back in so the new group applies. Membership of input lets any program you run read the keyboard; to undo it, run sudo gpasswd -d "$USER" input and delete the rule file.

Choose a speech model and dictate

  1. Pick a speech model in Models: download a local Whisper.cpp model, or paste a key for a cloud provider (AssemblyAI, Groq, Soniox, ElevenLabs or xAI). Parakeet needs an Apple Silicon Mac, so on Linux the local choice is Whisper.cpp.
  2. Set a shortcut, then hold it in any app and speak.
  3. Release it, and Entune pastes the text where you are.

Entune puts the text on the clipboard and on the primary selection, then presses Shift+Insert. Most apps paste the clipboard on Shift+Insert, and terminals such as GNOME Terminal, Konsole, xterm and kitty paste the primary selection, so the same keystroke reaches both. Linux offers no general way to confirm that text arrived, so Entune reports that the paste was sent. Every transcript is also kept in History.

On your desktop

  • Wayland: Entune’s window runs through XWayland, so the recording pill can show while another app has the focus, and wl-copy puts the text on the Wayland clipboard for the app you are in.
  • Tray: the tray item appears on KDE, and on GNOME with the AppIndicator extension, which Ubuntu includes. Without a tray, closing the window keeps Entune running for your shortcut; open it from the applications menu to bring the window back.

Linux support is newer: macOS has been tested by hand, and Windows and Linux are covered by automated tests. If something doesn’t work on your desktop, the issue tracker is the place to say so.

Go further

To keep everything on your computer, start with a local model and the optional processing off: set up local dictation. To get names and jargon right, build a personal dictionary: correct names and technical terms. The full setup reference is the Linux section of the guide.