Navigate a busy school corridor
Explore how turning speed and obstacle distance affect navigation. This sandbox does not model human collision risk.
Embodied AI · Robot control · Learn by doingMeet MayaGo.
Meet MayaGo.
Make intelligence move.
An original, self-hosted 3D robot lab. Drive, turn, test recovery and switch between walking legs and wheels—all inside your browser, without an external simulator.
MAI
SenseActLearn3D ENGINE READYMODE · WALKINGPOSE · STANDING
SPEED0.00 m/s*
HEADING000°
CONTROLIDLE
STABILITYSTABLE
MAYAGO / TRAINING ARENA SIMULATED WORLD
Welcome to MayaGo. Use WASD or the control pad to move.
MOVEMENT
WASD / Arrow keys
ROBOT ACTIONS
LIVE TELEMETRY
Forward command
Turn command
Gait phase
*Illustrative speed estimate. This learning build uses a custom kinematic motion model, not the original trained policy.
Uganda missions
Learn with a purpose.
Test a behaviour, record what your team learned, and identify what needs proper physics or real-world validation.
Recover after losing balance
Trigger the recovery sequence, repeat it and note its limits. An animated recovery is not proof of real-world reliability.
Plan a water-point inspection
Design a route and list the measurements, sensors and human checks needed for a useful inspection.
Turn an experiment into evidence.
Think like a robotics engineer.
- Observe: record what the robot actually did.
- Hypothesise: change one factor at a time.
- Measure: compare repeatable outcomes.
- Protect: define stop conditions and human oversight.
- Validate: real robot deployment requires real physics, trained policies and safety testing.
This version is a self-hosted 3D learning sandbox. It does not yet include MuJoCo WASM physics or the original ONNX policy weights.
Explore national challenges ↗