Robot learning frameworks

Open source

Lightweight interface for controlling ROS-based robotic arms using LeRobot

Open source

Community imitation-learning and RL workflows for a real SO-101 arm.

Open source

Linux process isolation using unshare, with ROS use cases.

Open source

Toolkit for commanding robots through mixed interaction modalities.

Open source

LIV-Eye: A Low-Cost LiDAR-Inertial-Visual Fusion 3D Sensor for Robotics and Embodied AI.

Open source

Robot manipulation policy research using consistency flow training.

Open source

Container format and tooling for pub/sub and robotics recordings.

Open source

MesaTask: Towards Task-Driven Tabletop Scene Generation via 3D Spatial Reasoning

Open source

ROS package for the Mini-Turty robot.

Open source

The ModAR world-action model implementation with RoboTwin training workflows.

Open source

Pub/sub messaging for Python, C and C++.

Open source

Performance-guided data-collection research for on-robot learning.

Open source

A touch-guided drawer-opening experiment with pi0.5 and Genesis.

Open source

An operating-system project connecting AI agents with physical hardware.

Open source

Physical AI Studio is an end-to-end framework for training robots to perform tasks through imitation learning from human demonstrations.

Open source

Piper text-to-speech integration for ROS 2.

Open source

A Python interface for MoveIt 2.

Open source

Robot-arm manipulation demonstration using the RAI framework.

Open source

Robot-description assets for the RBY1 mobile manipulator.

Open source

URDF descriptions for the Reachy robot.

Open source

A modular, primitive-first, python-first PyTorch library for Reinforcement Learning.

Open source

RoboBrain research repository for robotic manipulation reasoning.

Open source

Robot-agent harness with a CLI and local Web console for LIBERO short evaluation

Open source

Robot manipulation policy implementations for the Push-T task in LeRobot.