MicroDuck Simulator: Controls & Your First Walk
Try MicroDuck in your browser with the current keyboard controls, a first-walk exercise, reset steps and help for slow loading or unresponsive keys.
Read the guide →Practical explanations for software, simulation, training and the moments when something does not work.
Try MicroDuck in your browser with the current keyboard controls, a first-walk exercise, reset steps and help for slow loading or unresponsive keys.
Read the guide →Choose a MicroDuck route for Linux, Windows, Mac or ARM hardware. Separate browser play, CUDA training, local replay and cloud jobs before installing.
Read the guide →Read DuckEMW dance timeline source notes: tempo, first-beat offsets, movement segments and policy compatibility for MicroDuck simulation.
Read the guide →Interpret DuckEMW simulation speed results, including reported average and upright-gated peak speed, with source revisions and measurement limits.
Read the guide →Prepare the official MicroDuck RL environment with a pinned revision, Python 3.12, a CUDA check and a short training run before a longer experiment.
Read the guide →Diagnose MicroDuck simulator loading, ignored keys, Python errors, missing CUDA, ARM dependencies and ONNX input mismatches with source-linked checks.
Read the guide →Export a MicroDuck checkpoint with observation normalization, select the correct playback input layout and inspect a policy package before sharing.
Read the guide →Run the real MicroDuck daemons against a MuJoCo body, inspect health and motion with robotctl, measure real-time speed, and understand what this setup cannot test.
Read the guide →Inspect the policies on a MicroDuck, update the official set, try a local or community ONNX policy, and return one slot or the whole robot to official defaults.
Read the guide →Every guide carries its source, applicable version and verification status. Community results are kept distinct from our own tests.
See the editorial method →