What is the K-Bot platform?

K-Bot is the flagship 1.4m humanoid robot from K-Scale Labs, featuring 20+ degrees of freedom and designed for research and deployment.

What compute hardware is used onboard K-Bot?

The platform utilizes onboard Jetson or custom compute modules capable of running inference policies at 50-100 Hz.

What software stack is required for K-Bot deployment?

K-Bot typically uses the KSIM environment for training, KOS for firmware and actuator commands, and ONNX or JIT for policy export.

How does this skill support MuJoCo simulations?

The skill provides MJCF (MuJoCo XML) model configurations, including joint definitions, actuator Kp/Kd values, and mesh directories.

K-Bot Humanoid Robotics

Name: K-Bot Humanoid Robotics
Author: plurigrid

byplurigrid

•

Data Science & ML

Provides hardware specifications, MuJoCo MJCF models, and deployment workflows for the K-Scale flagship humanoid robot platform.

About

The K-Bot Humanoid skill is a comprehensive resource for developers working with K-Scale Labs' flagship robotics platform. It offers deep integration for managing hardware specifications, high-fidelity MuJoCo (MJCF) simulation models, and end-to-end deployment pipelines. By bridging the gap between high-level policy training in KSIM and real-world execution via KOS firmware, this skill enables seamless development of humanoid behaviors, from initial RL training to onboard Jetson-based inference and actuator control.

Key Features

2 GitHub stars
Streamline the deployment pipeline from KSIM training to KOS firmware
Access detailed hardware specs for the 1.4m, 20+ DOF humanoid platform
Generate and configure MJCF models for MuJoCo physics simulations
Standardize actuator configurations for position, velocity, and torque control
Integrate with the K-Scale ecosystem using validated GF(3) skill triads

Use Cases

Modeling robot kinematics and dynamics for simulation-to-real transfer
Developing reinforcement learning policies for humanoid locomotion in KSIM
Configuring real-time inference workflows for onboard Jetson compute modules

About

Key Features

2 GitHub stars
Streamline the deployment pipeline from KSIM training to KOS firmware
Access detailed hardware specs for the 1.4m, 20+ DOF humanoid platform
Generate and configure MJCF models for MuJoCo physics simulations
Standardize actuator configurations for position, velocity, and torque control
Integrate with the K-Scale ecosystem using validated GF(3) skill triads

Use Cases

Modeling robot kinematics and dynamics for simulation-to-real transfer
Developing reinforcement learning policies for humanoid locomotion in KSIM
Configuring real-time inference workflows for onboard Jetson compute modules