Download
To download the latest version of the application, click the policy checkbox below, then click the download button.
See Help for initial installation and usage instructions, or see Version Upgrade if you are upgrading from a previously installed version. Also see the License page to get a license key.
Release Notes 
Version 2.00 (2025-08-15)
- The application is now installed using an installer executable, so you no longer need to unzip the download file. When running the installer, you can select the installation directory and an option to create a desktop shortcut. The application is still a portable application, so it does not make any changes to your system outside of the installation directory and optional desktop shortcut.
- The application is now published with Microsoft, so warnings about the application being unpublished should go away. You may still see a warning when first starting the application, but this warning will eventually go away after the application has earned enough trust with Microsoft.
-
Created the new
repeat
function. This function allows you to define a sequence of actions (keyPress, sleep, type), where the overall sequence can be executed one or more times. - The alpha mode has been created to optimize typing of alphanumeric and other characters. This mode is useful for passwords, IDs, etc.
-
Created the new
openLink
function, which can be used in commands to open a URL in your default browser. See the new "open setup link" command in the "global" mode as an example. - Created a new "open chat" mode for OpenAI ChatGPT. Depending on the type of writing or coding you do, using generative AI tools like this can help you get things done faster, in fewer syllables. We are planning other mode-based integrations with generative AI tools in future versions.
- For the "windows" mode, created the commands: "open utterly folder", "insert utterly folder", "open startup folder", "open desktop folder", "open download folder", and "open calculator". The "list windows" exclusive mode also now lets you say a number to select a window.
- For the "calculator" mode, created the "close window" command.
- For the "chrome" mode, created the commands: "toggle developer tools" and "toggle developer console".
- For the "advanced" mode, changed the page up/down commands to "humvee/hummer".
- For the "global mode", renamed the "quit utterly voice" command to "close utterly".
- Improved the "emacs" mode to avoid accidental commands when using dired. This requires an update to your emacs configuration file. Also removed the "go start/finish" commands, renamed "quit emacs" to "close emacs", added "select line", added "select buffer", and added repetition utterance argument to "kill". Say "open help emacs" for details.
- Updated the Run Programs document to describe how to use the Windows Run Command.
- Improved the reliability of finding the settings files and log file when the application starts.
- Fixed bug that carried utterance forward when the microphone was shut off mid utterance.
Version 1.13 (2025-05-23)
-
You can now open the application window
as a normal, resizable, movable window.
See the
userInterfaceHeight
setting in the general settings file. -
Added the
darkTheme
setting to the general setting, which can be used to enable a dark theme for the primary user interface. -
Added the
microphoneInitiallyOn
setting to the general setting, which can be used to control whether the microphone is initially on or off. -
Added the "repeat last" command to the "basic" mode
and the related
repeatLast
function to the supported functions list. This command can be used to easily repeat any utterance one or more times. - Added three new modes that are useful for creating markdown documents, including technical documents. Added a new "markdown" mode for editing markdown content. Added a new "notable" mode for using the Notable application to create and edit markdown files. Added a new "latex" mode for creating LaTeX mathematical equations. See an overview of these in the Markdown and Mathematics documentation.
- Added a new "calculator" mode, which is used for the Windows calculator application. For details, say "open help calculator".
- Added a "symbols" mode, which can be used to type many special symbols. This mode is active by default.
- Added support for the "media-next-track", "media-previous-track", and "media-stop" keys.
- Improved the reliability of microphone connections for when microphones are disconnected.
- Added "new paragraph" command to the "basic" mode, which presses enter twice.
- Added "confirm", "confirm all", and "skip" commands to the "emacs" mode, and removed the "short yes" and "short no" commands. Also added the "extended command" command. Say "open help emacs" for details.
- Notable improvements to the YAML Tutorial.
- Added a new Troubleshooting page.
Version 1.12 (2025-03-03)
- The Microsoft Azure recognizer now supports custom speech models. This can be trained with your voice and with your jargon.
- Swaps are now supported to modify a recognized transcript before the interpreter processes the transcript for commands and typing. This feature is primarily useful for defining jargon and enforcing capitalization of certain words, but it can be used for other purposes. The "example jargon" and "basic" modes have been updated to use swaps.
- Macros are now supported. You can use this feature and the associated commands to record a sequence of utterances and replay them. You can use macros for typing repetitive content, clicking specific locations on the screen, issuing keyboard shortcuts, pausing input, typing long content that contains increasing numbers or letters, etc.
- Added the new execute function, which can be used to build commands that start applications and run command line executables. You can optionally send utterance words as flags to the executable and optionally type the results. With this feature, you have limitless ability to control what is typed when issuing commands.
-
Added the new
microphoneHotKey
general setting to define a Windows hot key that can be used to turn the microphone on and off. This is useful if you have a programmable foot pedal that you can use to emit the hot key. - Added new commands to the chrome mode: "toggle cursor", "open new window", "open private window", and "load archive". For more information, say "open help chrome".
- Many improvements to the "notepad plus plus" mode.
- Added "capslock" to the list of supported keys, and the "press caps lock" command to the "basic" mode.
- Added "forward email" command to both gmail modes, and renamed the "gmail default" mode to just "gmail".
- Added "upper case that" and "lower case that" commands to the "emacs" mode.
- Updated the Foot Pedals guide with new recommendations.
Version 1.11 (2025-01-25)
- Microsoft Azure is now supported as a recognizer. This is the first supported recognizer that provides locale-specific English language options. This should improve recognition quality for users with accents. We are also planning to add support for custom models with this recognizer in the next release.
- Added a configurable default value to the "number" function.
-
Added
initiallyActiveStates
setting to the general settings. This can be used to control which interpreter states are active when the application starts. - The "show" command now includes a "show multi" variant that repeatedly shows screen labels for clicking with brief pauses between each label showing. This is useful when you are navigating something that requires several successive left clicks. For more information, see the screen analyzer documentation.
- Added a new mode design guide to the documentation.
-
Added a new
audioFiles
setting to thesettings.yaml
file. If you want to verify the quality of your utterance audio, you can use this setting to create audio files of utterances. - Added new "zoom in", "zoom out", and "print page" commands to the "chrome" mode.
- Moved mouse commands from the "basic" mode to a new "mouse" mode, and moved interpreter state commands from the "basic" mode to the "global" mode. This makes it easier for users that do not use mouse commands to disable all mouse commands by deactivating the "mouse" mode.
Version 1.10 (2024-10-06)
- Multiple Monitors are now supported for all mouse control methods.
-
Added several new
keys
that can be used by the
keyPress
andkeyRepeat
command functions. -
Added a new
logRedaction
setting to thesettings.yaml
file. Set this field to true if you do not want dictation transcripts saved to the log file in the application directory (log.txt). Note that the mode and command names found in transcripts are still logged. - Improved reliability of "go back", "go forward", and "reload page" commands in the "chrome" mode. These commands were updated to use new browser-specific key presses. This allows you to use these commands even while focus is in an editable field.
- Added new media commands to the "basic" mode to play and pause music/videos and to adjust volume.
- Added new print screen commands to the "windows" mode for taking screenshots.
- Created the Chrome Optimizations help page with guidance for reducing popups while navigating.
Version 1.09 (2024-09-04)
- Created the new "reload settings" command to reload number and mode settings files while the application is running. This is useful to quickly test customizations. For more information, say "open help global".
- Changed the auto scrolling commands in the basic mode (and the associated function) to accept an optional number of seconds for scroll period. Say "open help basic" for details.
-
Modified the "title" command to not capitalize certain words
like "the", unless the word is the first or last argument to the command.
You can customize the word list with the
titleSkip
setting in theconfig\settings.yaml
file. - The application now detects when duplicate commands from multiple modes have become active, and it opens an error popup message.
- Fixed the user interface utterance history to not automatically scroll down when it is in focus.
- Fixed microphone button recovery when microphone is disconnected while the microphone is paused.
- Improved error handling when the application is installed in a directory that does not provide read and write file access.
- Improved reliability of "other window" command for Windows 11.
- Improved "show" and "mouse position" commands for high pixel density monitors.
- Created new foot pedals guide.
- Various improvements to documentation.
Version 1.08 (2024-08-15)
- Changed the keyboard shortcuts for the "chrome" mode Vimium commands: "show links", "show force links", and "show new links". The new keyboard shortcuts avoid conflicts with some popular web applications. If you are upgrading from a prior version of Utterly Voice, you will need to update your "custom key mappings" in Vimium settings to use F20 and F21 keys. See the updated Vimium documentation for instructions.
- Dropped support for Windows 8. This was necessary to support scaling changes as described below.
-
Decreased the default size of the user interface and provided a
userInterfaceHeight
general setting for custom sizes. -
Created a
labelFontHeight
general setting for the mouse commands ("show" and others) label font size. - Upgraded the application to handle scaling (DPI) changes for high pixel density monitors.
- Upgraded mode and command help to use local HTML files and the default web browser. This allows for easy printing of command help. Say "open help" or "open help basic" to try it out.
- Added the new commands-only interpreter state. When enabled, only commands are processed and normal typing is ignored. See the "Commands Only" section in the Interpreter State document.
- Added a new "context" command to the basic mode for typing a single word that is difficult to recognize without additional context. See the "Context" section of the Improve Recognition document.
- You can now click the microphone signal indicator to open the setup window. This is useful when your microphone settings are very far off and utterances are not being processed.
- Changed the "activate mode" and "deactivate mode" commands to just "activate" and "deactivate".
- Changed "type [key]" commands like "type home" to "press [key]" for improved recognition. The only exception to this is "type home/end" have been changed to "go home/end".
-
Created the new "advanced" mode for more efficient text editing.
Say "open help advanced" for command descriptions.
To activate this mode, say "activate advanced".
To automatically activate this mode when the application starts,
change the
initiallyActive
field in theconfig/modes/advanced.yaml
file to true. -
Renamed the following "basic" mode commands,
some of which have moved to "advanced":
- "quid" > "question mark"
- "bang" > "exclamation mark"
- "clear left" > "backspace"
- "clear right" > "delete"
- "open parent" > "open parentheses"
- "close parent" > "close parentheses"
- "undo that" > "scratch that"
- Simplified command settings by making "biasFactor" an optional field with a default value of 1.2.
- Modified the "escape" command to accept multiple words rather than just one word.
- Fixed number function to handle "0.0".
- Fixed the "lower" and "upper" commands so that they can be used to issue keyboard shortcuts manually while a modifier key is being held.
Older release notes
See release messages on the Google Group.