Skip to content
roman-16Public

About

Voice dictation for GNOME on Wayland. Press Super+Space, speak, and the words are typed into the focused field - transcribed live by Gemini or Mistral, with no daemon and nothing pasted.

Resources

Contributing

Security policy

Stars

3 stars

Watchers

1 watching

Forks

Repository files navigation

Murmur

Speak into any text field in GNOME, or record everything you hear and say.

Release GNOME Session License

Murmur transcribing a spoken sentence live and inserting it into a text editor

Press Super+Space, say what you mean, and the words appear where your cursor already is: in a browser, an editor, a chat box or a terminal. No window to switch to, no daemon in the background, and nothing to paste afterwards.

Press Super+Alt+Space instead and Murmur records everything your computer plays and your microphone hears - a call in Teams, Meet or Discord, a video, anything - until you press it again. Then it transcribes the lot, a line for everything said with when and by whom, and the audio and the transcript sit side by side in ~/Documents/Murmur, or the folder you choose.

  • You keep working while it listens. A panel the size of a notification appears at the bottom of the screen you are working on; it takes nothing over. Look anywhere else - another window, the overview, the window you were already in - and it collapses by itself, leaving an indicator with the countdown in the top bar. It is on screen exactly while it has your keyboard, so it is never in the way and never swallowing keys.
  • It lands where you are looking. Where there is a field, the transcription is typed into it, in any application, terminals included, with nothing pasted and your clipboard untouched. Which field is decided when you stop, so you can go and find it while you talk.
  • You watch it happen. A level that moves with your voice, a countdown, the destination named the whole time, and an optional hands-free stop after silence. With Grok your audio streams over a WebSocket as you speak and the words appear in the panel as they arrive; with Mistral Voxtral and OpenRouter the dictation goes up when you stop, to models made for whole files, and every transcription model OpenRouter serves is one key away.
  • Nothing is ever lost. If no text field is focused when you stop, Murmur copies the transcription to the clipboard instead of firing a sentence worth of keystrokes at whatever happens to be in front.
  • You can go back to what you said. Every dictation is kept on your machine and listed in the preferences, newest first: click one to copy it, search them, or clear the lot. A field the application reports as a password is never kept.
  • A recording stays out of the way for hours. It lives in the top bar alone, as a pill with how long it has run; its menu shows what the microphone and the desktop are hearing, and stops or discards it. It carries on while the screen is locked. When you stop it, a notification opens the transcript or copies its path, and a recording that could not be transcribed keeps its audio and asks again.
  • A transcript says who spoke and when. Your microphone and what the computer plays are kept on channels of their own, so every line you said carries your name, the voices on the call are numbered, and each line starts with the time of day it was said. Names and terms you list in the preferences are spelled the way you wrote them.
  • A transcript says where it was made and what was going on. It opens with the computer, the day and the times, the microphone and output, and the service that wrote it down. Among the lines said, it notes in brackets the window in front, what played, which application had the microphone, the camera, a shared screen and time away from the keyboard, so a call reads differently from a video, and one computer's recordings from another's. None of it goes to the service.
  • Any audio file, any time. Transcribe a file in the preferences sends a recording again, or a voice memo or a talk from anywhere, to the service you pick, and saves the transcript beside it.

Install

curl -fLo /tmp/murmur.zip https://github.com/roman-16/murmur/releases/latest/download/murmur@roman-16.github.io.shell-extension.zip && gnome-extensions install -f /tmp/murmur.zip && rm /tmp/murmur.zip

Then log out and back in, which Wayland requires for a new extension, and enable it:

gnome-extensions enable murmur@roman-16.github.io

You need GNOME Shell 47 to 51 on Wayland, GStreamer with its base and good plugins and PipeWire's GStreamer plugin, which GNOME's own screen recorder needs anyway, and an API key for one of the three transcription services. Installing with Nix, from source, from extensions.gnome.org once the listing is approved, updating and uninstalling: → Installation

Get started

Open the preferences, choose the service that transcribes for you, and paste in its key:

gnome-extensions prefs murmur@roman-16.github.io
Service Key from What it costs
Grok console.x.ai, once the account holds credit Dictation about $0.20 an hour of audio, recording about $0.20 an hour of a call where both sides talk. The words appear as you speak, a dictation runs as long as you set, and a recording as long as you like
Mistral Voxtral (default) console.mistral.ai Dictation about $0.24 an hour, recording about $0.50 an hour of a call where both sides talk, from the first minute. A dictation's words arrive when you stop; a dictation runs ten minutes at most
OpenRouter openrouter.ai/settings/keys One key for every transcription model it serves - Whisper, Nova, Voxtral, Parakeet - at each one's own price. A dictation's words arrive when you stop rather than as you speak; a dictation runs ten minutes at most

That is the whole setup. Put the cursor where the words belong, press Super+Space, and speak. Or press Super+Alt+Space to record.

Key What it does
Super+Space Start a dictation, and stop it the same way Enter does. Works from anywhere, always
Enter Stop and deliver the transcription
Ctrl+Enter Stop and copy to the clipboard
Esc Cancel, insert nothing, copy nothing

Enter, Ctrl+Enter and Esc reach the panel while it is on screen, which is exactly while it holds the keyboard. Look anywhere else and both go at once: the keyboard is yours again and the panel collapses to the top bar. Super+Space is a system shortcut and works either way.

Key What it does
Super+Alt+Space Start a recording of everything you hear and say, and stop it to have it transcribed. Click the pill in the top bar to watch the levels, stop it, or discard it

→ Getting started

Where your text goes

Murmur asks where the words go at the moment you stop, and the panel names the destination the whole time you are speaking, so it is never a surprise at the end.

What Murmur sees when you stop What it does
An application with a focused field Types the transcription into it with dotool, or the shell's virtual keyboard when dotool is unavailable
Nothing that can take text Copies it to the clipboard and says so, rather than turning your sentence into keyboard shortcuts

Because the question is asked at the end, you can start talking anywhere and click into the right field while you speak. Ctrl+Enter copies instead, whatever the panel says, for when the words belong somewhere other than the field in front of you.

Want to know whether an application is recognised? Turn on GNOME's on-screen keyboard and click into the field. If it pops up, Murmur sees that field too, because both read the same signal. → Where the text goes

What leaves your machine

Your voice, and while recording everything your computer plays, to the service you picked - api.mistral.ai by default, or api.x.ai or openrouter.ai - over a dictation or recording you started, along with the names and terms you listed. That is the only connection Murmur makes: no telemetry, no analytics, no update pings.

Dictation audio is never written to disk. The transcription is, on your machine only: the text of each dictation is kept in ~/.local/state/murmur@roman-16.github.io/history.jsonl so you can read it back later, until you clear it or turn Remember what I dictate off. A recording is written to disk whole: its audio and its transcript are saved side by side in the recordings folder, ~/Documents/Murmur unless you choose another, readable by you alone, and stay there until you delete them. The transcript notes the window titles on screen and what played during it, which Note what is on screen and playing turns off. Your API key is stored in dconf like every other GNOME setting, which means unencrypted, because extensions have no keyring access. If what you say must not reach a third party, Murmur is the wrong tool. → Privacy

Documentation

Page What's in it
Installation Every install route, requirements, updating, uninstalling
Getting started Your API key, your first dictation, the keys, your first recording
Configuration Every setting, what it changes, and its gsettings key
Where the text goes The insertion ladder, which apps are recognised, dotool setup
Troubleshooting Symptom, cause, fix
How it works Shortcut to text and shortcut to file, step by step, and what runs where
Privacy What leaves your machine, what is stored, what it costs
Limitations What Murmur will not do, and why

Good to know

  • A terminal is one big text field. It tells the compositor it accepts text whenever it is focused and nothing finer, so Murmur will happily deliver a sentence to vim in normal mode.
  • Dictation and recording are billed to your key. Murmur sends only what it records, and the ten-minute default keeps a forgotten dictation from running away. A recording runs until you stop it.
  • Tell people you are recording. A recording holds everyone in the call, and their voices go to the service too.
  • The voices on a call are told apart by the service. Your lines are yours because they came in on your microphone; the others share the desktop's channel, and the service numbers them by how they sound. A long recording goes up in ten-minute pieces, and someone who says nothing in the half minute around a cut comes back with a new number.
  • Over loudspeakers, the others are heard twice. Your microphone hears them as well, so their words also turn up under your name a moment later. Headphones keep the two apart.
  • Only Grok transcribes as you speak. Mistral and OpenRouter transcribe a dictation in one go when you stop, so with them the panel shows the level and the countdown while you talk, and the words all arrive at the end. Mistral does it with the model it uses for recordings, which also takes your names and terms; OpenRouter gives you the choice of every transcription model it serves, behind one key.
  • A dictation always arrives as one line. Line breaks are flattened to spaces before anything is typed, so a transcription can never press Enter in a chat box, a prompt or a shell. A tidied list keeps its bullets, inline.
  • The panel is not a window. It cannot be alt-tabbed or pushed behind an application, because a GNOME Shell extension draws inside the compositor rather than opening a window. Clicking anything else collapses it to the top bar instead, which a window behind a maximised application could not do.
  • The panel shows the last four lines. A long dictation scrolls past; the whole of it arrives in your field, and in the history, when you stop.
  • Wayland only. XWayland applications inside a Wayland session are fine; an X11 session is not.
  • dotool is recommended. Without it the fallback keyboard can only produce characters from your current layout, so emoji and other scripts are dropped.
  • Passwords are typed, never copied, never kept. A password field is still a text field, so nothing lands on the clipboard or in the history - as long as the application reports the field as one, which XWayland applications cannot.

Contributing

Issues, ideas and pull requests are welcome. CONTRIBUTING.md has the setup and the everyday commands, CHANGELOG.md records what each version changed, and SECURITY.md has the private channel for security reports.

License

MIT

About

Voice dictation for GNOME on Wayland. Press Super+Space, speak, and the words are typed into the focused field - transcribed live by Gemini or Mistral, with no daemon and nothing pasted.

Resources

Contributing

Security policy

Stars

3 stars

Watchers

1 watching

Forks

Releases

Packages

Used by

Contributors

Languages