Using ActionWright

Voice and Push-to-Talk

Dictate into any app with Push-to-Talk, or talk to ActionWright hands-free with Voice mode.

Two ways to use your voice

Push-to-TalkVoice mode
What it doesTypes what you say into the app you are using.Runs what you say as a task and speaks the answer.
How you start itHold ctrl+space.Say "Hey Yoo".
DefaultOnOff
MicrophoneOnly while you hold the keysListens in the background for the wake phrase
Needs a voice providerYes. See A voice provider.Yes

Before you begin

A voice provider

ActionWright turns your speech into text, and its replies into speech, through your voice provider. If your task provider has voice models, it is also your voice provider, with the same key, and there is nothing more to set up.

Some providers have no voice models, for example Anthropic at the time of writing. For those, Settings shows a Voice Provider section. Without a key there, Push-to-Talk and Voice mode cannot work, but typed tasks still run. See Task provider and voice provider.

Microphone access

  1. Open Windows Settings > Privacy & security > Microphone.
  2. Turn on Microphone access.
  3. Make sure apps are allowed to use the microphone, and that Action Wright is allowed if it is listed. Also turn on Let desktop apps access your microphone.
  4. ActionWright uses the Windows default microphone. There is no microphone choice in ActionWright, so set the default in Windows Settings > System > Sound.

Dictate with Push-to-Talk

Push-to-Talk is dictation. It works in any app with a text field, including the task window. It works whether Voice mode is on or off.

  1. Click where you want the text to go.
  2. Hold ctrl+space.
  3. Speak.
  4. Release the keys.
  5. Wait a moment. ActionWright transcribes your speech and inserts the text at the cursor.

You know it worked when the text appears. While you speak, a small indicator at the bottom of the screen shows bars that move with your voice. While ActionWright transcribes, it shows moving dots. The tray icon's tooltip also shows Listening…, Processing… and Done ✓.

Push-to-Talk and your clipboard

ActionWright inserts the text by pasting it. It puts the text on the clipboard and presses ctrl+v for you. It does not put back what was on the clipboard before, so after dictating, your clipboard holds the dictated text. If you need something you copied earlier, paste it before you dictate. If pasting fails, ActionWright types the text instead.

Good to know

  • The shortcut is fixed to ctrl+space. You cannot change it in Settings.
  • While ActionWright runs, all three of its shortcuts are reserved, so other apps don't receive them: ctrl+space, ctrl+shift+a and ctrl+esc. ctrl+esc is also the Windows shortcut for the Start menu.
  • If no text field has focus, nothing appears. The text is still on your clipboard, so you can paste it.
  • If Voice mode is on, Push-to-Talk comes first: pressing ctrl+space stops a spoken reply and stops Voice mode from listening.

The Push-to-Talk indicator

SettingDefaultEffect
Show a small indicator while using Push-to-TalkOnShows the small indicator at the bottom of the screen while you dictate. The change applies at once and does not affect Voice mode.

Open Settings from the tray icon (Settings) or from the task window (Menu > Settings).

Talk to ActionWright with Voice mode

Voice mode lets you start a task by saying "Hey Yoo" and hear the answer. Voice mode is off by default.

Turn on Voice mode

  1. Open Settings from the tray icon or from Menu in the task window.
  2. Tick Enable Voice mode (say "Hey Yoo" to start a task). The choice is saved when you tick it.
  3. Restart ActionWright: right-click the tray icon, choose Quit, and open ActionWright again. Voice mode starts only after a restart.

To turn Voice mode off, untick the box and restart ActionWright.

Give a task by voice

  1. Say "Hey Yoo".
  2. A short chime plays, and a small panel at the bottom of the screen shows ACTIVATED, then LISTENING....
  3. Say your request. Start within about 5 seconds, or ActionWright quietly stops listening.
  4. Stop talking. After a short pause, ActionWright stops recording and the panel shows PROCESSING....
  5. The task window opens and runs your request, like a typed task, in the current conversation.
  6. ActionWright speaks the answer and the panel shows SPEAKING.... The answer also appears under Result in the task window.

A long answer can take a moment before it starts, because ActionWright prepares all of the speech first.

Follow-up questions

After an answer is spoken, the panel shows LISTENING... again for about 10 seconds. Speak during that time to ask a follow-up, without saying "Hey Yoo" first. If you stay silent, the panel closes and ActionWright waits for the wake phrase again.

When a voice task needs you

If a task you started by voice asks a question, needs an approval, or pauses, ActionWright reads the question or prompt aloud. Answer it in the task window or the approval dialog. Saying "Hey Yoo" at that point stops the task instead. An approval dialog that is open stays open until you answer it. See When a task waits for you.

Voice tasks and the root folder

A voice request does not read the Root folder box. It uses the root folder from earlier in the same conversation, if a typed task already used one. Otherwise it runs without one, and for file work ActionWright may ask you to choose a folder in the task window. See Files.

The voice panel: Stop and New Chat

While Voice mode is active, a small panel at the bottom of your main screen shows what it is doing.

Panel showsWhat is happeningButtons
ACTIVATEDThe wake phrase was heard.None
LISTENING...ActionWright is recording your request or waiting for a follow-up.Stop, and New Chat during a follow-up
PROCESSING...Your request is being transcribed and run.Stop, New Chat
SPEAKING...The answer is being read aloud.Stop, New Chat
  • Stop ends what Voice mode is doing: it stops listening, stops a running voice task, or stops the spoken answer. Then ActionWright waits for "Hey Yoo" again.
  • New Chat does the same, and also starts a new conversation, like New Chat in the task window.

Interrupt a spoken answer

Do thisWhat happens
Say "Hey Yoo" clearly while it speaksThe answer stops, and ActionWright listens for a new request.
Say "Hey Yoo" clearly while a voice task is still runningThe task stops, and ActionWright listens for a new request.
Click Stop on the voice panelThe answer stops, and Voice mode waits for the wake phrase.
Press ctrl+escThe answer stops. If a task is running, it stops too.
Press ctrl+spaceThe answer stops, and Push-to-Talk starts recording.

What is listened to and what is sent

  • With Voice mode on, the microphone stays open in the background to listen for "Hey Yoo". Windows may show its microphone-in-use icon the whole time.
  • Listening for the wake phrase happens on your PC. That audio is not sent anywhere.
  • ActionWright records what you say after the chime, or during a follow-up, and sends that recording to your voice provider to be transcribed.
  • The transcribed text goes to your task provider as a task, like a typed one.
  • The answer text goes to your voice provider to be turned into speech.
  • With Push-to-Talk, only what you say while holding ctrl+space is recorded and sent.

Your providers' handling of this data is governed by their own terms. See Privacy.

Voice or the task window?

UseWhen
Push-to-TalkYou want to write by speaking: an email, a chat message, a note, or a task in the task box that you check before sending.
Voice modeYour hands are busy and the request is short: a quick question, a simple web lookup, a follow-up to the last answer.
The task windowThe request is long or exact, needs attachments or a root folder, or you want to follow the steps, answer questions and review the result.
  • Spoken words are easy to mishear. For names, numbers, addresses and file paths, type them, or dictate with Push-to-Talk and check the text first.
  • For tasks that change things, such as sending, buying or deleting, follow them in the task window, where you see every approval and step.

Troubleshooting

ProblemLikely causeWhat to try
Nothing happens when you say "Hey Yoo"Voice mode is off, or was turned on without a restart.Check the Voice mode box in Settings, then quit and reopen ActionWright.
Voice mode still does not react after a restartActionWright cannot use the microphone.Check microphone access and the Windows default microphone, then restart ActionWright.
The panel shows PROCESSING..., then closes with no answerYour speech could not be transcribed.Check the voice provider and its key in Settings, then try again.
Transcription failed notification after Push-to-TalkNo voice provider key, a provider problem, or nothing was recorded.Check the voice provider in Settings. Hold ctrl+space while you speak and release after you finish.
Dictated text does not appearNo text field had focus, or the app blocks pasting.Click in the field and try again. The text is on your clipboard, so you can paste it with ctrl+v.
Something you copied earlier is gonePush-to-Talk replaces the clipboard with the dictated text.Paste what you need before dictating.
The answer appears but is not spokenSpeech could not be generated by your voice provider.Check the voice provider and its key in Settings. If you entered your own STT or TTS model names under Voice Advanced Settings, clear them to use the current defaults.
Another app's ctrl+space, ctrl+shift+a or ctrl+esc shortcut stopped workingActionWright reserves its shortcuts while it runs. See Good to know.Use that app's menu instead, or quit ActionWright while you need the shortcut.

For more, see Troubleshooting.