Beta Testing¶

Thank you for joining the Murmure beta program! Your feedback is invaluable to make the application rock-solid before its official release.

How to Get the Beta¶

Beta builds are published before each release. Head over to the GitHub Releases page and download the latest pre-release version.

Murmure now uses Parakeet ultra, a fine-tuned version of Parakeet made by Moondream, which makes fewer mistakes
The model is the int8 export by @thiswillbeyourgithub, which uses about 380 MB less RAM than the previous one

The overlay can stay on screen after a dictation, with the text and a copy button. Choose when in Settings > System > Show result after dictation (Off, Command, Command and LLM, All dictations), and choose how long it stays
Move the mouse over the result to keep it on screen

New Automatic insert switch in Settings > System, separate from the insertion method. When it is off, Murmure never types the text by itself. The text stays in the history, and in the clipboard if Copy to Clipboard is on
The Paste last transcript shortcut always pastes, even when Automatic insert is off
If you used the insertion method "None (manual paste)", it is converted for you. Automatic insert is off and your other settings are kept

LLM Connect is now called Prompt Mode
Command Mode is a separate extension, with its own model. Enable it in Extensions > Command Mode
With nothing selected, a command now answers your question instead of repeating it

The local API is no longer experimental. It has its own page in Extensions > Local API
New endpoints apply a Prompt Mode prompt to the transcription and return the final text
Each endpoint has a Try it button to test it with a WAV file, without writing any code

Hesitation sound removal ("euh", "um") is now optional and off by default, because it also removed real words such as the German "um". Turn it back on in Personalize > Formatting Rules
The tray menu is translated into your language
New --quit command line flag to close Murmure
Transcriptions no longer contain a stray <unk> token
macOS: clicking the Dock icon reopens the window
Linux: the volume of audio started during a dictation, such as a new browser tab, is restored at the end
Updates downloaded as .deb keep the right file extension
Several security improvements

Do what you can, even one box helps. Start with the four essentials, they take about five minutes.

[ ] Dictate a few sentences like you normally do, and check the text lands correctly and is accurate
[ ] In Settings > System, set Show result after dictation to All dictations, dictate, and check the result appears with a copy button, then disappears after the chosen duration
[ ] Turn off Automatic insert, dictate, and check that nothing is typed but the result still appears. Then use the Paste last transcript shortcut and check the text is pasted
[ ] If you updated from an older version, check that your settings, shortcuts and dictionary are still there

[ ] If you use Prompt Mode or Command Mode: enable Command Mode, then press the Command shortcut with nothing selected and ask a question. Then select a sentence and say "translate to English"
[ ] Open Extensions > Local API, enable it, and use Try it with a short WAV file
[ ] Dictate something long, over a minute, and check nothing is missing at the end
[ ] Open the tray menu and check it is in your language

[ ] macOS: close the Murmure window, then click the Dock icon, and check the window comes back
[ ] Linux: turn on the volume reduction in Settings > System, start a video in a new browser tab while you dictate, and check its volume comes back at the end
[ ] Linux: run murmure --quit in a terminal and check Murmure closes

No need to open a GitHub issue, just reply in the beta announcement conversation. Tell us what broke and on which OS, that is already enough.

If you can, add the steps to reproduce it and the log file (enable debug mode in Settings > System, then reproduce the bug).

Thank you for your contribution!