Skip to main content
Pre-release

This release is a pre-release and may not be stable for production use.

Kuavo Humanoid SDK

A comprehensive Python SDK for controlling Kuavo humanoid robots. This SDK provides interfaces for robot state management, arm and head control, and end-effector operations. It is designed to work with ROS (Robot Operating System) environments.

Warning: This SDK currently only supports ROS1. ROS2 support is not available.

Warning: This SDK can only be used on the onboard NUC computer located in the robot's torso.

Kuavo 4Pro Robot

Features

  • Robot State Management

    • IMU data (acceleration, angular velocity, euler angles)
    • Joint/motor states (position, velocity, torque)
    • Torso state (position, orientation, velocity)
    • Odometry information
    • End-effector states:
      • Gripper(lejuclaw): position, velocity, torque, grasp status
      • Dexterous hand(qiangnao): position, velocity, torque
      • Touch Dexterous hand(qiangnao_touch): position, velocity, torque, touch state
      • End-effector position and orientation
    • Motion states: stand, walk, step_control, trot
  • Motion Control

    • Arm Control
      • Joint position control
      • End-effector 6D control via inverse kinematics
      • Forward kinematics (FK) for computing end-effector pose
      • Keyframe sequence control for complex motions
    • End-effector Control
      • Gripper control (position control with configurable velocity and torque)
      • Dexterous hand control
        • Position control
        • Pre-defined hand gestures (OK, 666, fist, etc.)
    • Head Control
      • Position control
    • Torso Control
      • Height control (squatting)
      • Forward/backward tilt control
    • Dynamic Motion Control
      • Stance
      • Trot
      • Walking (xy and yaw velocity control)
      • Stepping (gait switching)
  • Robot Basic Information

    • Robot type (kuavo)
    • Robot version
    • End-effector type
    • Joint names
    • Total degrees of freedom (28)
    • Arm degrees of freedom (7 per arm)
    • Head degrees of freedom (2)
    • Leg degrees of freedom (12)

Installation

Note: There are currently two versions of this SDK, the stable version and the beta version. Their differences are:

  • stable version: corresponding to the functionality provided by the master branch of kuavo-ros-opensource.
  • Beta version: This version is more aggressive than the official version and also provides richer functionality, corresponding to the functionality provided by the beta branch of kuavo-ros-opensource.

Friendly reminder: Please be clear about which version you need to install. If your SDK version does not match kuavo-ros-opensource, some features may not be available.

Install the latest ​stable version of Kuavo Humanoid SDK using pip:

pip install kuavo-humanoid-sdk-ws

Install the latest ​beta version of Kuavo Humanoid SDK using pip:

pip install --pre kuavo-humanoid-sdk-ws

For local development installation (editable mode), use:

cd src/kuavo_humanoid_sdk_ws  
chmod +x install.sh  
./install.sh   

Upgrade Instructions

Before upgrading, you can check the currently installed version with:

pip show kuavo-humanoid-sdk-ws
# Output:  
Name: kuavo-humanoid-sdk-ws  
Version: 0.1.2  
...  

Note: If the version number contains the letter b, it indicates a beta version, e.g., Version: 0.1.2b113

To upgrade from a stable version to the latest stable version:

pip install --upgrade kuavo-humanoid-sdk-ws

To upgrade from a beta version to the latest stable version:

pip install --upgrade --force-reinstall kuavo-humanoid-sdk-ws
# or  
pip uninstall kuavo-humanoid-sdk-ws && pip install kuavo-humanoid-sdk-ws

To upgrade from a stable/beta version to the latest beta version:

pip install --upgrade --pre kuavo-humanoid-sdk-ws

Package Information

You can check the package information using pip:

pip show kuavo-humanoid-sdk-ws

Quick Start

Here's a simple example to get started with Kuavo Humanoid SDK:

Warning: Before running any code, make sure to start the robot first by executing either:

  • For simulation: roslaunch humanoid_controllers load_kuavo_mujoco_sim.launch (Example command)
  • For real robot: roslaunch humanoid_controllers load_kuavo_real.launch (Example command)
# Copyright (c) 2025 Leju Robotics. Licensed under the MIT License.
import time
from kuavo_humanoid_sdk import KuavoSDK, KuavoRobot

def main():
    if not KuavoSDK().Init():  # Init! !!! IMPORTANT !!!
        print("Init KuavoSDK failed, exit!")
        exit(1)
    robot = KuavoRobot()    
    
    """ arm reset """
    print("Switching to arm reset mode...")
    robot.arm_reset()
    
    """ stance """
    print("Switching to stance mode...")
    robot.stance()

    """ trot """
    print("Switching to trot mode...")
    robot.trot()
    
    """ walk forward """
    print("Starting forward walk...")
    duration = 4.0  # seconds
    speed = 0.3     # m/s
    start_time = time.time()
    while (time.time() - start_time < duration):
        robot.walk(linear_x=speed, linear_y=0.0, angular_z=0.0)
        time.sleep(0.1)  # Small sleep to prevent busy loop
    
if __name__ == "__main__":
    main()

Docs

The documentation is available in two formats:

We recommend that you generate the documentation locally by running the documentation script. The documentation will be output to the docs/html and docs/markdown folders:

cd <kuavo-ros-opensource>/src/kuavo_humanoid_sdk_ws
chmod +x gen_docs.sh
./gen_docs.sh

We recommend that you view the documentation using html for a better experience.

For Markdown documentation at:

https://gitee.com/leju-robot/kuavo-ros-opensource/tree/master/src/kuavo_humanoid_sdk/docs/markdown/index.md

Examples

WARNING

Before running any code examples, make sure to start the robot first by executing either:

  • For simulation: roslaunch humanoid_controllers load_kuavo_mujoco_sim.launch (Example command)
  • For real robot: roslaunch humanoid_controllers load_kuavo_real.launch (Example command)

Robot Info

Examples showing how to get basic robot information.

https://gitee.com/leju-robot/kuavo-ros-opensource/tree/master/src/kuavo_humanoid_sdk/examples/robot_info_example.py

Basic Robot Control

A basic example showing how to initialize the SDK and control the robot’s movement.

https://gitee.com/leju-robot/kuavo-ros-opensource/tree/master/src/kuavo_humanoid_sdk/examples/motion_example.py

End Effector Control

LejuClaw Gripper

Examples demonstrating how to control the LejuClaw gripper end effector, including position, velocity and torque control.

https://gitee.com/leju-robot/kuavo-ros-opensource/tree/master/src/kuavo_humanoid_sdk/examples/lejuclaw_example.py

QiangNao DexHand

Examples showing how to control the QiangNao DexHand, a dexterous robotic hand with multiple degrees of freedom for complex manipulation tasks.

https://gitee.com/leju-robot/kuavo-ros-opensource/tree/master/src/kuavo_humanoid_sdk/examples/dexhand_example.py

Arm Control

Examples showing arm trajectory control and target pose control.

https://gitee.com/leju-robot/kuavo-ros-opensource/tree/master/src/kuavo_humanoid_sdk/examples/ctrl_arm_example.py

Forward and Inverse Kinematics

Examples demonstrating how to use forward kinematics (FK) to compute end-effector positions from joint angles, and inverse kinematics (IK) to calculate joint angles needed to achieve desired end-effector poses.

https://gitee.com/leju-robot/kuavo-ros-opensource/tree/master/src/kuavo_humanoid_sdk/examples/arm_ik_example.py

Head Control

Examples showing how to control the robot’s head movements, including nodding (pitch) and shaking (yaw) motions.

https://gitee.com/leju-robot/kuavo-ros-opensource/tree/master/src/kuavo_humanoid_sdk/examples/ctrl_head_example.py

Step-by-Step Control

Examples showing how to control the robot’s movements step by step, including individual foot placement and trajectory control.

https://gitee.com/leju-robot/kuavo-ros-opensource/tree/master/src/kuavo_humanoid_sdk/examples/step_control_example.py

License

This project is licensed under the MIT License - see the LICENSE file for details.

Contact & Support

For any questions, support, or bug reports, please contact:

Metadata

Release files for kuavo-humanoid-sdk-ws 1.4.4b1473

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for kuavo-humanoid-sdk-ws 1.4.4b1473
File Size Uploaded Provenance Actions
kuavo_humanoid_sdk_ws-1.4.4b1473.tar.gz 144.9 kB Without provenance Details for kuavo_humanoid_sdk_ws-1.4.4b1473.tar.gz

Built distributions (wheels)

Table of built distributions (wheels) for kuavo-humanoid-sdk-ws 1.4.4b1473
File Interpreter ABI Platform Actions
kuavo_humanoid_sdk_ws-1.4.4b1473-20260630114640-py3-none-any.whl Python 3 none any Details for kuavo_humanoid_sdk_ws-1.4.4b1473-20260630114640-py3-none-any.whl
kuavo_humanoid_sdk_ws-1.4.4b1473-20260630082302-py3-none-any.whl Python 3 none any Details for kuavo_humanoid_sdk_ws-1.4.4b1473-20260630082302-py3-none-any.whl
kuavo_humanoid_sdk_ws-1.4.4b1473-20260629212220-py3-none-any.whl Python 3 none any Details for kuavo_humanoid_sdk_ws-1.4.4b1473-20260629212220-py3-none-any.whl

Total release size: 658.0 kB

Release files / kuavo_humanoid_sdk_ws-1.4.4b1473.tar.gz

Download URL kuavo_humanoid_sdk_ws-1.4.4b1473.tar.gz
Size 144.9 kB
Tags Source
SHA-256 checksum
How to use checksums
a230c685fcf71a439595693a53fa554ceb2b02f9471898d8e76459478c48b048
BLAKE2b-256 checksum
How to use checksums
34fd392a008ed826e977b7a1a49a6b43c097870b12d2f94396ac78a074cba50f
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/6.1.0 CPython/3.8.10

Release files / kuavo_humanoid_sdk_ws-1.4.4b1473-20260630114640-py3-none-any.whl

Download URL kuavo_humanoid_sdk_ws-1.4.4b1473-20260630114640-py3-none-any.whl
Size 171.0 kB
Tags Python 3
SHA-256 checksum
How to use checksums
6a113caa74c79fe639d3b5851199749c6fb6da67ad2aafae0376c95e2c7599cf
BLAKE2b-256 checksum
How to use checksums
59b81bca8f053b171b0ae470dd5d42027ae026319035d8a07b8c24cac868db0c
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/6.1.0 CPython/3.8.10

Release files / kuavo_humanoid_sdk_ws-1.4.4b1473-20260630082302-py3-none-any.whl

Download URL kuavo_humanoid_sdk_ws-1.4.4b1473-20260630082302-py3-none-any.whl
Size 171.0 kB
Tags Python 3
SHA-256 checksum
How to use checksums
9fca5fbdd74992d51ae3af387a6139f1f2f6e85f63c5d970d851e6279f5b8ad6
BLAKE2b-256 checksum
How to use checksums
56f1427e8eb797a31118820d66e1271b40f6037dd1b853b490f140e52a2230b1
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/6.1.0 CPython/3.8.10

Release files / kuavo_humanoid_sdk_ws-1.4.4b1473-20260629212220-py3-none-any.whl

Download URL kuavo_humanoid_sdk_ws-1.4.4b1473-20260629212220-py3-none-any.whl
Size 171.0 kB
Tags Python 3
SHA-256 checksum
How to use checksums
ff9106ad308dbed3d422e922f82cd83c7ba10525d268398ba9454c7f3d42effa
BLAKE2b-256 checksum
How to use checksums
de4793cbee57ba7d6ced9e5c51b0436ee147726099fe6c61e945fd3d40541269
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/6.1.0 CPython/3.8.10

Security for kuavo-humanoid-sdk-ws 1.4.4b1473

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

No provenance found

None of the files in this release were uploaded with provenance metadata.

A note on package security

PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.

To learn more about protecting yourself from malicious actors, visit our package security documentation.

Release history Release notifications | RSS feed

This release
Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page