Hands-free works with laptop speakers: while the robot talks it only listens for “stop”, so it never answers its own voice. Speech recognition and voices come from your browser; Chrome works best.
Each turn is timed from the moment you stop talking (or press send) to the moment the voice starts.
Short notes the robot keeps on this device. They go with each turn and never leave this browser otherwise.
Call mode listens all the time. With speakers, while the robot talks it only listens for “stop” or 「停」, then starts listening afresh, so it never answers its own voice. With headphones, set INTERRUPT BY to talking over it. A sentence you finish after a short pause joins the one before it.
Speech recognition and voices are your browser's. Chrome works best. Camera frames and your words go to the brain for one turn and are not stored on our server.