Skip to content

LeRobot

An open robotics project sharing models, datasets and simulation environments

Visit official site

What LeRobot is

The repository describes LeRobot as an open project intended to lower the barrier to working in robotics, so that people can contribute to and benefit from shared datasets and pretrained models. It says the included approaches are validated in the real world, with a focus on imitation learning and reinforcement learning. Pretrained checkpoints, demonstration datasets collected by people, and simulation environments are provided so experiments can start without a physical robot.

What you can do with it

  • Study manipulation with the ACT policy in the ALOHA environment
  • Simulate arm control with SimXArm and TDMPC
  • Explore pushing tasks with PushT and a diffusion policy

Who it is for

  • Researchers and developers, per the source
  • Educators and students using it as a teaching tool
  • Robotics hobbyists exploring personal projects

What to watch out for

  • Support for affordable real robots is described as planned future work

Pros & cons

✓ What we like

  • Concrete policy and environment pairings are listed
  • Datasets, checkpoints and simulators are bundled together
  • Setup documentation and visualisation tools are described

! What to watch out for

  • Real-robot support is described as planned, not shipped
  • Python environment and dependency work is required
  • Per-model and per-dataset licences are outside this description

FAQ

Can it be used without a physical robot?

Simulation environments are provided so experiments can start without assembling hardware.

Which learning approaches are included?

The repository highlights imitation learning and reinforcement learning.

Does it support real robots now?

The post describes support for affordable real robots as planned future work.

Last reviewed: 2026-11-15

More AI coding tools tools

View all →

How we review