EmbodiK

High-performance inverse kinematics solver optimized for cross-embodiment VLA/AI applications

EmbodiK is a modern C++ library with Python bindings designed for robust, high-performance IK behaviors in cross-embodiment scenarios. The name reflects its focus on embodied dynamics and constraint handling, making it ideal for humanoid robots and AI/VLA integrations. Built on top of Pinocchio and using Nanobind for seamless Python integration.

Features

🚀 High Performance: C++ core with optimized Eigen linear algebra
🐍 Python Integration: Seamless numpy array support via Nanobind
🎯 Multiple Solvers: Single-step and full multi-task velocity IK
🛡️ Singularity Robust: Advanced inverse methods for stable solutions
🔒 Constraint Support: Joint limits and operational space constraints
📊 Visualization: Optional Viser-based interactive visualization

Quick Start

import embodik
import numpy as np

# Create robot model
model = embodik.RobotModel.from_urdf("path/to/robot.urdf")

# Create kinematics solver
solver = embodik.KinematicsSolver(model)

# Solve IK for a target pose
target_pose = np.eye(4)  # 4x4 transformation matrix
result = solver.solve_position_ik(target_pose)

if result.status == embodik.SolverStatus.SUCCESS:
    print(f"Solution: {result.solution}")

Installation

See the Installation Guide for detailed instructions.

pip install embodik

Documentation

Installation Guide — How to install EmbodiK
Quickstart — Get started in 5 minutes
GPU Solvers — FI-PeSNS and PPH-SNS GPU-accelerated solvers
API Reference — Complete API documentation
Examples — Example code and tutorials
Development Guide — Contributing and development
Recovery robustness — Hardware seeds, limits, collision, stall recovery

Preview

Franka Panda collision-free IK

ROBOTIS AI Worker constraint teleop

License

Apache License 2.0 - see the LICENSE file for details. Source: robodreamer/embodik