Prepare speech recognition
Voice commands act in place of buttons in the Vision workspace. The same Vision and Safety prerequisites apply.
- Open Settings > Speech recognition.
- Install the local speech model requested by the app.
- Check microphone input.
- Hold V or Hold to speak, wait until the app shows that it is listening, then say one phrase at a time.
Command examples
Set the app language to English and use these phrases.
| Purpose | Command |
|---|---|
| Select target | select red cube / select green cube / select blue cube |
| Clear target | clear target |
| Start arm following | start following |
| Stop arm following | stop following |
| Start object detection | start object detection |
| Stop object detection | stop object detection |
Voice commands follow the same safety checks. They do not automatically enable torque or bypass an emergency stop.
Check the recognition result and actual app state. Do not rely on a voice response alone to confirm that the robot has stopped or torque has been released.
Language and input behavior
Speak in the app's selected language. To use another language, change the app language before recording.
Release V or the button to submit. The V shortcut does not work in text input fields or modal dialogs. If you release it before recording is ready, the audio may not be sent. Leaving or hiding the window during recording cancels it.
Selecting a target does not move the arm or enable torque. Start following starts detection if it is off, or requests arm following if detection is active and a target is selected. Check the displayed result. Use the command phrases in the table.
Stop following stops arm following and leaves detection active. Stop object detection stops both arm following and detection. The alias stop tracking also stops both. Neither command releases torque during a normal stop. Support the arm before turning Servo torque off.