English | 中文
LimX Dynamics is a general-purpose robotics company, focusing on full-size humanoid robots and have developed innovative products such as biped and quadruped robots.
LimX Dynamics is dedicated to disruptive innovations by Embodied AI. The goal is to unleash the generalization of Artificial General Intelligence (AGI) in the physical world. Based on revolutionizing core software and hardware technologies, and building first-of-its-kind foundation model for humanoid robots, we aim to achieving loco-manipulation robotics platform and Embodied AI tool kits for innovators and integrators. We promote the widespread applications of Embodied AI across B2B & B2C sectors, including R&D, manufacturing, business and households services.
Open source projects
| Platform | Type | Title | Description | Stars | Forks |
|---|---|---|---|---|---|
| TRON2 | Description | tron2-robot-description | Robot model files for TRON2 variants, including URDF/xacro, MuJoCo XML, meshes, and optional USD assets for simulation, visualization, and downstream tooling. |  |
 |
| Simulation | tron2_gazebo_ros | A ROS Noetic and Gazebo simulation workspace for TRON2 robots that supports controller integration, Sim-to-Real validation, and deployment workflows. |  |
 |
|
| Simulation | tron2_mujoco_sim | A MuJoCo simulation tool for TRON2 robots that supports SF/WF robot types and controller deployment through the LimX SDK. |  |
 |
|
| Training | tron2_rl_lab | An Isaac Lab reinforcement learning training stack for TRON2 robots, supporting SF/WF (sole-foot and wheel-foot) variants. |  |
 |
|
| Training | TRON2_YG_LAB | An Isaac Lab reinforcement learning training stack for TRON2 robots with 6-DoF arm and gripper variants, designed for locomotion policy training and play. |  |
 |
|
| Deployment | tron2_rl_deploy_python | A Python-based reinforcement learning deployment demo for TRON2 robots that runs ONNX policies and supports MuJoCo simulation and real-world deployment. |  |
 |
|
| Deployment | tron2_rl_deploy_ros | A ROS Noetic reinforcement learning deployment workspace for TRON2 robots, including hardware, controller, and ONNX runtime packages for simulation and real robots. |  |
 |
|
| Deployment | tron2_env | A public TRON2 runtime package for OpenPI deployment, providing WebSocket robot communication, motion execution, observation collection, and RTC helper utilities for real-robot manipulation. |  |
 |
|
| Deployment | tron2_openpi | A TRON2 deployment-focused derivative of OpenPI that adds TRON2 policy transforms, deployment configuration templates, pi0/pi0.5 policy serving, and real-robot client examples. |  |
 |
|
| Manipulation | tron2_manipulation_controller | A local simulation and SDK validation workspace for the TRON2 dual-arm robot, pairing a MuJoCo motor-level simulator with the controller runtime and WebSocket SDK, and including VLA trajectory replay and limit checking. |  |
 |
|
| Navigation | tron2-navigation-ros2 | ROS 2 Humble navigation integration for TRON2, including the production SCAN-Planner subset, RoboSense LiDAR drivers, and safety/deployment tooling; GPL components live in companion repos. |  |
 |
|
| Navigation | tron2-navigation-pct-gpl | GPL-2.0-or-later PCT planner sources for tron2-navigation-ros2. |  |
 |
|
| Navigation | tron2-navigation-fastlio-gpl | GPL-2.0-only FAST-LIO ROS 2 sources for tron2-navigation-ros2. |  |
 |
|
| Oli | Description | humanoid-description | Robot model files, including URDF, USD, and XML model files. Can be used for simulation, training, etc. |  |
 |
| Simulation | humanoid-mujoco-sim | A MuJoCo simulation tool for the Oli robot that supports quick Sim-to-Real validation and deployment of robot algorithms. |  |
 |
|
| Training | humanoid-rl-isaaclab | A reinforcement learning training framework designed specifically for Oli humanoid robots, built on Isaac Lab , providing comprehensive tools for simulation and training, and supporting sim-to-real locomotion. |  |
 |
|
| Deployment | humanoid-rl-deploy-python | A Python-based reinforcement learning deployment algorithm that streamlines the process of deploying trained models on Oli robots. |  |
 |
|
| Deployment | humanoid-rl-deploy-cpp | A lightweight C++ algorithm framework that enables rapid deployment of trained models when ROS1/ROS2 is not required. |  |
 |
|
| Deployment | humanoid-rl-deploy-ros2 | A reinforcement learning deployment framework based on ROS2 that enables rapid deployment of trained models on Oli robots. |  |
 |
|
| Deployment | humanoid-rl-deploy-ros | A reinforcement learning deployment framework based on ROS1 that enables rapid deployment of trained models on Oli robots. |  |
 |
|
| Luna | Description | luna-description | Robot model files for the LimX Luna humanoid — URDF, MuJoCo MJCF, and NVIDIA USD (closed-loop variant) for simulation, visualization, and downstream tooling. |  |
 |
| Algorithm | luna-beyondmimic | A BeyondMimic-based whole-body motion learning framework that teaches the LimX Luna humanoid to dance — from human dance motions to deployable policies on Isaac Lab. |  |
 |
|
| Algorithm | GMR | LimX Dynamics fork of GMR (General Motion Retargeting) — retarget human motion (SMPL-X/BVH/Xsens/OptiTrack) into LimX humanoids (OLI EDU, Luna) and 17+ other robots in real time on CPU. |  |
 |
|
| TRON1 | Description | tron1-robot-description | Robot model files for TRON1 variants, including URDF/xacro, MuJoCo XML, and meshes for point-foot and wheel-legged robots. |  |
 |
| Simulation | tron1-gazebo-ros | A ROS1-based simulation tool for the TRON1 robot that supports quick Sim-to-Real validation and deployment of robot algorithms. |  |
 |
|
| Simulation | tron1-gazebo-ros2 | A ROS2-based simulation tool for the TRON1 robot that supports quick Sim-to-Real validation and deployment of robot algorithms. |  |
 |
|
| Simulation | tron1-mujoco-sim | A MuJoCo simulation tool for the TRON1 robot that supports quick Sim-to-Real validation and deployment of robot algorithms. |  |
 |
|
| Training | tron1-rl-isaaclab | A reinforcement learning training framework designed specifically for TRON1 robots, built on Isaac Lab , providing comprehensive tools for simulation and training, and supporting reinforcement learning in various environments. |  |
 |
|
| Training | tron1-rl-isaacgym | A reinforcement learning training framework designed specifically for TRON1 robots, built on Isaac Gym , providing comprehensive tools for simulation and training, and supporting reinforcement learning in various environments. |  |
 |
|
| Deployment | tron1-rl-deploy-ros | A reinforcement learning deployment framework based on ROS1 that helps you quickly deploy your trained models. |  |
 |
|
| Deployment | tron1-rl-deploy-ros2 | A reinforcement learning deployment framework based on ROS2 that helps you quickly deploy your trained models. |  |
 |
|
| Deployment | tron1-rl-deploy-python | A reinforcement learning deployment algorithm that leverages Python to streamline the deployment of your trained models. |  |
 |
|
| Deployment | tron1-rl-deploy-arm | A reinforcement learning deployment framework for TRON1 arm controllers based on ONNX runtime. |  |
 |
|
| Navigation | tron1-ss | A ROS Noetic-based SLAM and autonomous navigation stack for TRON1 robots, supporting LiDAR mapping, localization, and path planning. |  |
 |
|
| Agent | tron1-agent | Real-time voice interaction agent for Tron robots with LLM-powered dialogue, speech recognition (FunASR), TTS, tool calling, and task management. |  |
 |
|
| VLA | Algorithm | FluxVLA | All-in-one VLA engineering platform for embodied intelligence — from data collection to real-robot deployment. |  |
 |
| Simulation | FluxBisim | IsaacSim-based bimanual manipulation simulation for FluxVLA. |  |
 |
|
| Algorithm | FluxDAgger | A model-decoupled DAgger pipeline for dual-arm robotic manipulation — human-in-the-loop data collection compatible with arbitrary VLA and reward models. |  |
 |
|
| SDK & Tools | SDK | limxsdk-lowlevel | Low-level SDK for direct robot communication (motors, sensors, real-time control), used by the simulation and deployment repos above. |  |
 |
| Visualization | robot-visualization | Real-time visualization tool for robot joint states, IMU, force sensors, and motion trajectories via RViz/PlotJuggler. |  |
 |
|
| Bridge | ros1-bridger | ROS1 communication bridge (x86_64 & aarch64, Noetic) for connecting development machines to LimX robot networks. |  |
 |
|
| Bridge | ros2-bridger | ROS2 communication bridge (x86_64 & aarch64, Foxy/Humble/Jazzy) for connecting development machines to LimX robot networks. |  |
 |
|
| Tool | robot-joystick | Cross-platform virtual joystick for simulation and real-robot manual control. |  |
 |
|
| CLI | limx-cli | CLI tool and Scratch visual programming bridge for LimX robots, with JSON output for AI Agent integration. |  |
 |
|
| CLI | gradmotion-cli | CLI tool for GradMotion whole-body motion generation on LimX robots with AI Agent integration support. |  |
 |
|
| TronCamp | Event | troncamp-mani | TronCamp ACT 4-task contestant kit — self-contained package (Tron2 dual-arm / RoboTwin / single-GPU ACT) with setup, starter, and eval for four increasingly difficult manipulation tasks. |  |
 |
| Web | troncamp-web-mani | TronCamp hackathon public site with task briefs, T1-T4 freedom ladder, in-process evaluation, and token-anonymous leaderboard. |  |
 |
|
| Web | troncamp-web-locomotion | TronCamp locomotion track web companion for the TRON2 robot. |  |
 |
|
| Profile | Profile | limxdynamics | GitHub profile page for LimX Dynamics — this README overview plus auto-updating total-star / forks growth charts. |  |
 |