Linux how-to · Facts verified 2026-09-01

How to do speech to text on Linux
community tools and WhisperJot beta

GNOME and KDE do not ship built-in dictation today. Community tools such as nerd-dictation and Speech Note can transcribe offline, Fedora 42 offers ibus-speech-to-text, and Canonical has announced Myna for Ubuntu 26.10 but it is not released. WhisperJot for Linux is in beta: a hotkey captures speech and inserts text via wtype, ydotool, or xdotool, with a clipboard fallback.

Get WhisperJot Try the browser demo
Built-in path

How to do Linux dictation without a built-in tool

There is no built-in dictation in GNOME or KDE today. The options below are community projects, a Fedora change, and an announced Ubuntu feature that is not shipping yet.

  1. Confirm there is no desktop built-in. GNOME and KDE do not include a system dictation feature today. Plan on a community tool, a distribution extra, or a third-party app.
  2. On Fedora 42, try ibus-speech-to-text. Fedora 42 offers ibus-speech-to-text, which uses VOSK via IBus. That is a distribution change, not a GNOME or KDE built-in that every Linux desktop has.
  3. Try nerd-dictation for offline VOSK. nerd-dictation runs offline with VOSK. It types via xdotool on X11, and via ydotool, wtype, or dotool on Wayland.
  4. Or use Speech Note for files and live text. Speech Note transcribes offline with Whisper or Vosk, installs as a Flatpak, and on Wayland needs the ydotool daemon to insert text.
  5. Account for how text is injected. xdotool is X11-only. wtype depends on compositor support. ydotool works everywhere via uinput but needs its daemon and access to /dev/uinput.
  6. Do not wait on Ubuntu Myna just yet. Canonical is building Myna for Ubuntu 26.10: local models, hotkey-triggered, Wayland and GNOME first. It was announced in June 2026 and is not released.
Limits to know

Limits to know

  • No built-in dictation ships in GNOME or KDE today.
  • Ubuntu Myna is announced for 26.10 and is not released.
  • Text injection depends on the display server: xdotool will not help on Wayland, and ydotool needs a daemon plus /dev/uinput access.
  • Speech Note on Wayland needs the ydotool daemon before it can insert text.
The WhisperJot way

How to dictate on Linux with WhisperJot (beta)

  1. Install the Linux beta package. Download the .deb (Debian and Ubuntu are the most-tested path) or the tar.gz for another modern 64-bit distribution. WhisperJot for Linux is in beta.
  2. Bind a hotkey your desktop will honor. The hotkey can come from the desktop portal on Wayland, from evdev if your user is in the input group, or from a CLI socket. Press it with the cursor in the app that should receive text.
  3. Let WhisperJot insert the transcript. Text is inserted with wtype, ydotool, or xdotool, with a clipboard fallback when those tools cannot type into the focused field.
  4. Use local engines by default. Jot Local and Jot Local Pro transcribe on-device and offline after the model download. Jot Cloud is opt-in, runs only while selected, and audio is never stored on our servers. Sessions cap at 30 minutes.
Tips

Make the first session count.

Match the tool to the display server

xdotool is X11-only. On Wayland, plan on wtype (compositor support) or ydotool (uinput daemon). WhisperJot tries those, then clipboard fallback.

Fedora and Ubuntu are not the same

Fedora 42's ibus-speech-to-text is a distro extra. Ubuntu's Myna is announced for 26.10 and not released. Neither is a GNOME or KDE built-in.

Treat WhisperJot Linux as beta

The .deb on Debian or Ubuntu is the most-tested install. Portal, evdev, and CLI-socket hotkeys exist because Linux desktops do not share one input API.

Need a file transcribed?

Speech Note transcribes files offline. WhisperJot can also import an audio file of up to 60 minutes from the menu bar with the local engines.

Questions

Straight answers.

Is there built-in Linux dictation?

Not in GNOME or KDE today. Fedora 42 offers ibus-speech-to-text (VOSK via IBus) as a distribution change. Canonical announced Myna for Ubuntu 26.10 in June 2026, with local models and a hotkey, Wayland and GNOME first, but it is not released. Until a desktop ships a built-in, Linux dictation means community tools or a third-party app.

What speech to text tools work on Linux?

nerd-dictation runs offline with VOSK and types through xdotool on X11 or ydotool, wtype, or dotool on Wayland. Speech Note is a Flatpak that uses Whisper or Vosk offline and needs the ydotool daemon to insert text on Wayland. Fedora 42 has ibus-speech-to-text. WhisperJot Linux is a beta desktop app with a global hotkey and clipboard fallback.

How does WhisperJot work on Linux?

The Linux build is in beta as a .deb or tar.gz. Capture a hotkey through the desktop portal on Wayland, evdev when your user is in the input group, or a CLI socket. WhisperJot transcribes with local engines by default and inserts text with wtype, ydotool, or xdotool, falling back to the clipboard. Jot Cloud is opt-in and used only while selected.

What is Ubuntu Myna?

Myna is Canonical's announced speech-to-text feature for Ubuntu desktop. It is planned around local models, a hotkey, and Wayland plus GNOME first, targeting Ubuntu 26.10. The announcement was in June 2026; it is not released, so it is not a tool you can turn on in current Ubuntu. Use community tools or WhisperJot's Linux beta until it ships.

Sources

Vendor documentation.

Sources, checked 2026-09-01: Ubuntu Discourse: Introducing Myna, Canonical Myna repository, Fedora: ibus-speech-to-text, nerd-dictation, Speech Note (dsnote), wtype, ydotool. Vendor features change — check each product's documentation for current behavior.

Related: Linux speech to text · App-by-app dictation guides · How to convert a voice recording to text · System requirements · 30-second browser demo

One hotkey, wherever you write.

WhisperJot types into the focused app and runs local transcription on your device by default.

Get WhisperJot — $99/year