Voice Trigger

Use spoken phrases to request supported app actions

Voice Trigger listens for phrases that you define and turns a recognized phrase into an app action. For example, a phrase can request a supported action such as Shoot when that action is available for voice.

Voice Trigger uses Windows speech recognition on this PC. It uses the Windows default microphone and the language of the signed-in Windows user.

Before you start

  • Make sure Windows can use your microphone.
  • Allow the app to access the microphone in Windows privacy settings.
  • Confirm Windows speech recognition or dictation works with the language you want to speak.
  • If you want to speak a different language, install the matching Windows language pack and sign in with that language or locale.
  • Choose short phrases that are easy to say and hard to confuse with normal conversation.

All speech recognition for this page is intended to run locally through Windows speech recognition.

Page controls

  • Start begins continuous listening.
  • Stop ends continuous listening.
  • PTT listens briefly for one phrase.
  • Minimum confidence controls how certain recognition must be before an action can run.
  • Actions enabled allows recognized phrases to run actions.
  • Actions disabled lets you test recognition without running actions.
  • The action picklist chooses what the phrase should do.
  • The spoken phrase box is where you type the phrase to recognize.
  • Save adds or updates the phrase/action mapping.
  • Edit loads the selected mapping for changes.
  • Delete removes the selected mapping.

The phrase/action list shows each saved mapping. The status history shows listening state, heard phrases, blocked actions, and requested actions.

Quick setup

  1. Open Voice Trigger from Triggers, Voice.
  2. Select an action from the action picklist.
  3. Type a short spoken phrase.
  4. Press Save.
  5. Leave Actions disabled while testing the phrase.
  6. Press PTT and say the phrase.
  7. Watch the status history for what the app heard and the confidence value.
  8. Adjust the phrase or Minimum confidence until recognition is reliable.
  9. Turn Actions enabled on when you are ready for recognized phrases to run actions.

Start with one phrase and one action. After that works, add more mappings one at a time.

PTT mode

PTT means push to talk. Press PTT when you want the app to listen for a phrase for a few seconds.

Use PTT when:

  • You only need voice triggering occasionally.
  • You are testing a new phrase.
  • The room is noisy.
  • You want the least chance of an accidental action.

You do not need to hold the PTT button down. Click it once, say the phrase, and then check the status history.

PTT can also be assigned to a hotkey on the Hotkeys page. That lets you press a key, speak a phrase, and keep your mouse away from the Voice Trigger page.

Continuous mode

Continuous mode keeps listening until you press Stop.

Use continuous mode when:

  • Your hands are busy.
  • You need repeated voice control during a session.
  • The room is quiet enough for reliable recognition.

For continuous mode, use phrases that are less likely to appear in normal conversation. For example, a phrase such as computer capture is safer than capture.

Continuous mode should be stopped when you no longer need it.

Profile startup

Voice Trigger remembers whether continuous listening was active when the profile was saved or when the app closed. If continuous listening was active, Voice Trigger starts continuous listening automatically the next time the selected profile is applied.

Stop Voice Trigger before saving the profile or closing the app when you want it to remain off the next time the profile is applied. This keeps speech recognition startup and shutdown cleanup out of normal app startup when Voice Trigger is not in use.

PTT is a momentary listening mode. Using PTT does not make Voice Trigger start automatically on the next app run.

Choosing phrases

Good phrases are short, distinct, and easy for Windows speech recognition to hear.

Use phrases such as:

  • computer capture
  • take picture
  • capture now

Avoid phrases that are:

  • Too short, such as go.
  • Similar to each other.
  • Common in nearby conversation.
  • Hard to pronounce consistently.
  • Mixed across languages.

The phrase is your data. It is not translated by the app. If you use a translated app UI, still type the phrase exactly as you want to speak it.

Minimum confidence

Minimum confidence is the required recognition confidence before a phrase can request an action.

Raise the value if the app hears the wrong phrase or triggers too easily. Lower the value only if Windows hears the correct phrase but the action is blocked for low confidence.

Use the status history while tuning. It shows the heard phrase and confidence percentage.

Actions enabled

Keep Actions disabled while setting up or testing. In this mode, the app can still listen and show what it heard, but it does not run the mapped action.

Turn Actions enabled on only after the phrase is recognized reliably.

This is especially important for actions that can affect camera state, capture images, or interrupt your workflow.

Action mappings

Each mapping connects one spoken phrase to one action. The action list contains only actions that the current app exposes to Voice Trigger.

Not every hotkey action is available for voice. Some actions are keyboard-oriented and are intentionally not exposed to Voice Trigger.

To change a mapping:

  1. Select the mapping in the phrase/action list.
  2. Press Edit or double-click the mapping.
  3. Change the action or phrase.
  4. Press Save.

To remove a mapping, select it and press Delete.

Status history

The status history is the first place to check when voice control does not behave as expected.

It can show:

  • Ready to configure voice phrases.
  • PTT listening.
  • Listening continuously.
  • Heard phrase and confidence.
  • Actions disabled.
  • Confidence below the minimum.
  • Action requested.
  • Recognition or microphone errors.

The main app Log may contain lower-level diagnostic details, but the Voice Trigger status history is the user-facing result.

Troubleshooting

If PTT or Start is disabled:

  • Add at least one phrase/action mapping.
  • Make sure the phrase is not blank.
  • Stop continuous listening before editing mappings.

If the app does not hear anything:

  • Check Windows microphone privacy settings.
  • Confirm the microphone is the Windows default input device.
  • Test the microphone with Windows dictation or another speech tool.
  • Close other apps that may be using the microphone.
  • Restart the app after changing Windows microphone permissions.

If the app hears the wrong phrase:

  • Use a more distinct phrase.
  • Avoid phrases that sound alike.
  • Add a prefix word for continuous mode.
  • Speak at a steady pace.
  • Reduce background noise.

If the phrase is heard but no action runs:

  • Turn Actions enabled on.
  • Check the status history for low confidence.
  • Lower Minimum confidence only after confirming the phrase is heard correctly.
  • Confirm the mapping uses the action you intended.
  • Confirm that action is available for Voice Trigger in this app.

If recognition fails or reports a language issue:

  • Confirm the Windows speech language matches the language you are speaking.
  • Install the needed Windows language pack.
  • Sign in with the matching Windows language or locale when needed.
  • Recreate the phrase using the words you will actually speak in that language.

If actions trigger by accident:

  • Turn Actions enabled on only after testing.
  • Raise Minimum confidence.
  • Use longer or more distinct phrases.
  • Use PTT instead of continuous mode.
  • Stop continuous listening when you are done.

Related setup

Use Hotkeys when you want keyboard-triggered actions. Use Scanner when barcode input should fill fields or trigger an action.

Voice Trigger, Hotkeys, and Scanner all request actions through the same action system, but each trigger source exposes only the actions that make sense for that source.