Robot
Agent Control
Drive the robot with your voice — hold Shift, speak a command.
Every session ships with a built-in voice agent. Hold Shift, speak, release. The agent transcribes your voice, interprets the intent, moves the robot, and replies verbally. It joins the session automatically when you start the robot — no extra setup.
For the pipeline behind this, see Concepts → Agent.
How to use
- Start the robot from its settings page.
- Open the cockpit view.
- Hold Shift — the mic pill at the bottom shows Listening… with an animated halo.
- Speak a command, then release Shift.
Example commands
| You say | What happens |
|---|---|
| "Move forward" | Robot walks forward; agent confirms |
| "Turn left a bit" | Robot rotates; agent confirms |
| "Stop" | Robot halts |
| "What do you see?" | Agent describes the robot's camera feed |
The agent is an LLM, so conversational phrasing works — "back up", "face the other way" — not just keywords.
The pipeline (speech-to-text → LLM → text-to-speech) introduces a short delay. For instant, deterministic control, use Manual Control.
How is this guide?