Open-Source AI · Robotics & embodied AI

LeRobot vs Diffusion Policy

LeRobot vs Diffusion Policy compared for 2026 — features, license, ease of use, performance and which one to choose. Train real robots with the Hugging Face stack vs Teach a robot by showing it, using diffusion.

Updated regularly · curated by OpenSourceAI.tech

Choose LeRobot for anyone teaching a physical robot new skills without building the plumbing first. Choose Diffusion Policy for cloning a demonstrated skill rather than engineering a controller.

LeRobot vs Diffusion Policy at a glance

SpecLeRobotDiffusion Policy
CategoryRobotics & embodied AIRobotics & embodied AI
TypeRobot learning libraryImitation learning
LicenseApache-2.0MIT
Runs locallyYesYes
Primary languagePythonPython
Ease of useIntermediateAdvanced
Best foranyone teaching a physical robot new skills without building the plumbing firstcloning a demonstrated skill rather than engineering a controller
GitHub stars4.4k

How LeRobot and Diffusion Policy score

🏆 Overall edge: LeRobot — 4.5 vs 3.4 / 5
CriterionLeRobotDiffusion Policy
Popularityn/a2.5
Maintenancen/a2.0
Ease of use3.52.5
Privacy5.05.0
License freedom5.05.0

Scores are computed automatically from public signals — GitHub stars (popularity), recent commit activity (maintenance), license type (freedom), local-first design (privacy) and onboarding complexity (ease of use). Indicative, not a verdict.

What each one is

LeRobot

Robot learning library · Apache-2.0

LeRobot brings pretrained models, datasets and simulation environments for real-world robotics into one Python library — the closest thing the field has to a Transformers moment.

  • Pretrained policies and shared datasets on the Hub
  • Works with cheap off-the-shelf arms, not just lab hardware
  • Backed by an active community and weekly releases
Visit LeRobot →

Diffusion Policy

Imitation learning · MIT

Diffusion Policy generates robot actions with a diffusion model — the technique that made visuomotor imitation learning finally work reliably.

  • State-of-the-art results on manipulation
  • Reference implementation from the original paper
  • Widely reused as a baseline
See the Diffusion Policy page →

Key differences

LeRobot is robot learning library, while Diffusion Policy is imitation learning. Their licenses differ (Apache-2.0 vs MIT), which matters if you ship a commercial product. LeRobot leans more intermediate-friendly, whereas Diffusion Policy is more suited to advanced users. In short, LeRobot fits anyone teaching a physical robot new skills without building the plumbing first, and Diffusion Policy fits cloning a demonstrated skill rather than engineering a controller.

Which should you choose?

Choose LeRobot for anyone teaching a physical robot new skills without building the plumbing first. Choose Diffusion Policy for cloning a demonstrated skill rather than engineering a controller.

There is rarely one winner — many setups use both. The right pick depends on your hardware, your team's skills, and whether you value simplicity or control.

Frequently asked questions

Is LeRobot or Diffusion Policy easier to use?

LeRobot is generally the easier of the two to get started with, while Diffusion Policy rewards more setup with more control.

Are LeRobot and Diffusion Policy free?

LeRobot is free and open source (Apache-2.0), and Diffusion Policy is free and open source (MIT). Neither charges for the core software.

Can I run LeRobot and Diffusion Policy locally?

LeRobot: yes · Diffusion Policy: yes. Both can be used without sending your data to a third-party cloud where their setup allows.

LeRobot vs Diffusion Policy — which should I pick in 2026?

Choose LeRobot for anyone teaching a physical robot new skills without building the plumbing first. Choose Diffusion Policy for cloning a demonstrated skill rather than engineering a controller.

People also compare

Explore more open-source AI

Browse thousands of open-source AI tools, models and projects — all curated in one place, updated daily.

Explore the directory →