Ross ROSS = Recommend OSS · open-source software intelligence for agents

huggingface/lerobot

🤗 LeRobot: Making AI for Robotics more accessible with end-to-end learning observed · 2026-08-28

github.com/huggingface/lerobot · homepage · Python · Apache-2.0 (permissive) observed · 2026-08-28

Health v2 · maintenance only

92/100

  • Activity 99
  • Release rhythm 96
  • Longevity 67
How is this computed?

round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10) — computed 2026-09-02. Adoption (stars, forks) is never an input.

  • gap_med: 28.5
  • age_days: 950
  • days_rel: 30
  • days_push: 7
  • n_releases_24m: 11

Full methodology

Adoption not part of the score

26949 stars · 5472 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

LeRobot is Hugging Face's PyTorch library for real-world robotics, providing a hardware-agnostic Python interface for controlling robots, a standardized dataset format (LeRobotDataset), and state-of-the-art imitation-learning and vision-language-action policies. It includes CLI tools and a GUI (LeLab) covering the full workflow: teleoperate, record, train, and deploy policies on real robots.

Use cases

  • train imitation learning policies for real robot arms
  • record and share robot demonstration datasets
  • control low-cost robot arms like SO-100 from Python
  • deploy vision-language-action models like pi0 on real robots
  • teleoperate a robot with a leader arm, keyboard, or gamepad
  • integrate custom robot hardware into a standard control interface
  • stream and visualize large robotics datasets from the Hugging Face Hub

When to choose

  • you want an end-to-end open-source stack for real-world robot learning
  • you use supported hardware like SO-ARM101, Koch, LeKiwi, or Reachy
  • you need a standardized dataset format for robot demonstrations
  • you want pretrained policies (ACT, pi0, SmolVLA) ready to train and deploy
  • you're a beginner looking for a low-barrier entry into physical AI

When to avoid

  • you only need simulation-only robotics without real hardware
  • your robot uses unsupported motors or communication protocols and you can't implement the Robot interface
  • you need hard real-time control guarantees
  • you require macOS support for video decoding via TorchCodec, which is unavailable there

Facets

library · maturity active

machine-learning deep-learning robotics sdk cli data-science robotics machine-learning deep-learning artificial-intelligence python windows cli robot-learning imitation-learning teleoperation vision-language-action-models robot-datasets pytorch hardware-agnostic policy-training huggingface-hub real-world-robotics datasets linux macos gpu

10 sources

Member repositories

RepositoryRoleHealth v2
huggingface/lerobotmain92

For agents

markdown · JSON · MCP: product_card(name="huggingface/lerobot")

Data as of 2026-08-30T08:39:29.467469+00:00 · Report a problem