Run a mock robotics benchmark and review evidence with Inspect Robots
Run the built-in CubePick benchmark with a scripted policy, inspect resolved configuration and scores, and render saved robotics evaluation logs for review.
Prerequisites
Python 3.11+; uv or equivalent virtual environment tooling; inspect-robots CLI
Installation
Install or set up from the source-backed instructions:
Use Python 3.11+ in a fresh project directory. Create a virtual environment with uv venv, then install the verified PyPI release with uv pip install inspect-robots==1.0.0 and activate .venv. Run inspect-robots list, then inspect-robots run --task cubepick-reach --policy scripted --embodiment cubepick. Inspect and render the emitted LOG.json using inspect-robots inspect LOG.json and inspect-robots view LOG.json. No robot plugins, GPU, Rerun extra or model key are needed for this bounded mock workflow. Documentation was inspected at an upstream source revision; packaged-release parity was not runtime-tested.