Multimodal perception
Combine camera vision, voice, distance, light, motion, and robot telemetry into one grounded view of the moment.
Z Autobot / model / Z2
A small robot
with local reflexes.
Z2 combines realtime conversation with physical awareness. It can listen, see, speak, move, sense its surroundings, adapt to ambient light, protect itself at an edge, and report what actually happened.
01 / Intelligence
Z2 is not a remote-controlled hardware kit with a chatbot attached. Its AI connects perception, reasoning, memory, tools, and physical execution into one robot system.
Combine camera vision, voice, distance, light, motion, and robot telemetry into one grounded view of the moment.
Understand intent, choose a safe plan, select the right tools, and adapt the response to the person and situation.
Turn decisions into bounded robot actions, then return the actual command status instead of assuming success.
Use managed realtime AI services for speech, vision, reasoning, memory, and continuously improving capabilities.
Recognize authorized people, preserve preferences and relationships, and maintain useful context across interactions.
Connect every machine to an account with transparent token allowance, consumption, and capability access.
02 / Toolbox system
Toolboxes give the AI explicit capabilities. Z2 can combine several toolboxes in one plan while permissions, local safety, and execution acknowledgements remain visible.
Play favorites, choose a mood, control playback, and use music inside routines.
Create, schedule, pause, resume, and report multi-step robot tasks.
Read schedules, create reminders, prepare for events, and coordinate daily routines.
Use a phone as a live controller, camera view, microphone, display, and robot companion interface.
Move toward a destination using local ranging, vision, safe-path decisions, and guarded recovery.
Support stories, routines, learning, reminders, and family check-ins while keeping adult supervision essential.
Deliver messages, start calls, announce updates, and help people stay connected.
Operate authorized lights, appliances, scenes, and connected devices through explicit permissions.
Explain, demonstrate, quiz, translate, and adapt educational activities to the person.
Remember names, preferences, recurring needs, important facts, and the context behind a request.
Identify objects, read visible information, describe scenes, and use visual context during a task.
Control motion, lights, display, sound, sensors, and presentation with local safety always in authority.
03 / Cloud system
04 / Hardware
Every component has a defined job in the robot: motion, awareness, communication, feedback, or protection.
Dual-core control with 16 MB flash and 8 MB PSRAM.
Differential two-wheel drive with measured motion calibration.
Front-left, front-right, rear, and ultrasonic front coverage.
Independent left and right cliff detection.
Forward visual input for image understanding and scene awareness.
Measures surrounding light for awareness and automatic interface brightness.
Digital microphone and amplified 3 W speaker path.
128 × 64 status display and ambient-responsive light ring.
A modern physical connection for power, setup, and wired maintenance.
05 / Awareness
Z2 does not wait for a remote service before responding to a nearby obstacle or a missing floor.
Front sonar, three directional ranging sensors, and two floor sensors continuously describe the space around Z2.
Obstacle and cliff rules run on the robot. A cloud connection is never required for the immediate stop boundary.
A guarded reverse or turn is allowed only when the corresponding path is known to be clear.
06 / Conversation
07 / Physical specification
Z2 / unique machine identity