Robot learning frameworks

Open source

Training a humanoid robot for locomotion using Reinforcement Learning

Open source

Publishes robot link transforms from joint states and robot descriptions.

Research

Examples for deploying ROS and ROS 2 workloads on Kubernetes and KubeEdge.

Open source

Modular reinforcement-learning tools for control and simulation workflows.

Research

Research code for injecting robotic affordances into multimodal language models.

Open source

ROS packages for the ABB YuMi robot.

Open source

Tools for experimenting with low-cost AI robots.

Research

A URDF model of the Robotiq 2-Finger Adaptive Gripper 85.

Open source

Official repository for "OMG: Omni-Modal Motion Generation for Generalist Humanoid Control", https://arxiv.org/abs/2606.10340.

Open source

A C++ behavior-tree library for task coordination.

Open source

Model packages for the Walkman robot demonstration.

Open source

Python package for robot learning.

Research

Research code for reinforcement learning with learned skill priors.

Research

Implementations of RVT and RVT-2 robot manipulation policies.

Open source

Research implementation of an open-ended language-model agent in Minecraft.

Open source

A benchmark for language-grounded robot manipulation in realistic 3D scenes.

Open source

AutoEval: Autonomous Evaluation of Generalist Robot Manipulation Policies in the Real World | CoRL 2025

Open source

Official Codebase for "Do as I Do: Dexterous Manipulation Data from Everyday Human Videos"

Open source

Video-Action Models for Generalizable Robot Control Beyond VLAs

Research

URDF, SDF, and mesh assets for robotics projects.

Open source

Motion-capture data collection tools for dexterous robot manipulation.

Open source

A Rust robot runtime for deterministic execution and replay.

Research

Tools for reading Dex-Net object, grasp and grasp-quality data.

Research

URDF and mesh assets for the Agility Robotics Digit robot.