From one LLM call to a squad of robots taking orders in plain English. Every concept comes with a simulation you can play and the real Python code you can copy and run.
Everyone codes with AI assistants now, so this course is about concepts and capabilities: what agents can do and which piece solves which problem. The simulations run in your browser with no API key, and the code is real, so your robot moves for real when you run it.
Pick a request, press Run and watch the robot think, call tools and read its manuals, message by message.
Every demo maps to a Python script in code/. Same tools, same world, same results, with a real LLM behind it.
Each chapter builds on the last: R-80 gets a brain, then hands, a manual, a memory and finally a squad.
A lab robot controlled by an LLM agent. Give it an order.
How we use LLM agents as high-level controllers of real robots at the BioRobotics Group: capability manifests, a safety gate the LLM can't bypass, and a slow reasoning loop on top of fast classic control.
Open the field chapter ▶Each stage adds one piece to the same robots. By the end, they coordinate a mission on their own.