mink
Python inverse kinematics based on MuJoCo
What it is and what it does
mink is a Python library for solving inverse kinematics (IK) problems on robot manipulators using the MuJoCo physics engine. It computes joint angles that move a robot's end-effector to a desired position and orientation while respecting joint limits, velocity bounds, and collision constraints. The library supports both single-arm and multi-arm systems, dexterous hands, humanoids, and legged robots, and can handle closed-chain kinematics through equality constraints.
The package depends on mujoco as its core physics backend, qpsolvers for quadratic programming, and typing_extensions for type hints. It provides a Lie group interface for rigid body transformations with native C optimization for performance-critical IK computations. Installation is straightforward across Linux, macOS, and Windows with pre-built wheels, though you will need a valid MuJoCo model and understanding of task specification to use it effectively.
Use it for:
- Real-time motion planning for robotic arms to reach target poses while avoiding obstacles.
- Humanoid robot control where multiple kinematic chains and balance constraints must be satisfied simultaneously.
- Dexterous manipulation tasks requiring hand and arm coordination with collision avoidance.
- Legged robot locomotion planning with closed-chain constraints from ground contact.
- Research prototyping of inverse kinematics algorithms on diverse robot morphologies.
Worth the install?
AI-flagged interpretation of the facts on this page — verify before relying
mink solves inverse kinematics problems for robot manipulators using MuJoCo, supporting task specification in configuration or operational space with collision avoidance and closed-chain constraints.
Yes, if you are working with MuJoCo-based robot simulation or control. mink is actively maintained, has no known vulnerabilities, uses a permissive license, and solves a well-defined problem (differential IK) with broad robot support. Install friction is moderate but manageable. Not suitable if you need IK for non-MuJoCo simulators or if your robot morphology is not covered by existing examples.
Install
mink on PyPI
pip
pip install minkuv
uv add minkpoetry
poetry add minkInstalling mink
Before you install
Medium install friction due to compiled wheels across multiple platforms (cp310–cp312, macOS, Linux, Windows, musllinux). Active maintenance with recent release (55 days old) and 1493 repository stars suggest solid upkeep.
License in practice
Apache-2.0 is permissive; you may use, modify, and distribute mink freely in commercial and private projects, provided you retain license notices.
Quickstart
pip install mink
import mink
# Requires MuJoCo model; see documentation for task setup
# solver = mink.IKSolver(...)
Requires MuJoCo physics engine and a valid robot model file; Python 3.10–3.13.
Verify before relying
- Whether qpsolvers and typing_extensions are always required at runtime or only for specific features.
- Performance characteristics and scalability limits for real-time control loops.
- Compatibility with MuJoCo versions beyond what the fact sheet specifies.
Package facts
| License | Apache-2.0 (permissive) |
| Python support | supports the current Python release (<3.14,>=3.10) |
| Install friction | medium — platform-specific wheel |
| Runtime dependencies | 3 — mujoco, qpsolvers, typing_extensions |
| Maintenance | actively maintained — 55 days since the last release |
| Last repo commit | |
| First released | |
| Downloads | 337,911/month — #7,442 on PyPI (30-day window, as of 2026-08-14) |
| Known vulnerabilities | none known (OSV.dev, checked 2026-08-14) |
Evidence: mink-1.2.0-cp310-cp310-macosx_10_9_x86_64.whl; mink-1.2.0-cp310-cp310-macosx_11_0_arm64.whl; mink-1.2.0-cp310-cp310-manylinux1_x86_64.manylinux_2_28_x86_64.manylinux_2_5_x86_64.whl; mink-1.2.0-cp310-cp310-manylinux2014_aarch64.manylinux_2_17_aarch64.manylinux_2_28_aarch64.whl; mink-1.2.0-cp310-cp310-musllinux_1_2_aarch64.whl; mink-1.2.0-cp310-cp310-musllinux_1_2_x86_64.whl; mink-1.2.0-cp310-cp310-win32.whl; mink-1.2.0-cp310-cp310-win_amd64.whl; mink-1.2.0-cp311-cp311-macosx_10_9_x86_64.whl; mink-1.2.0-cp311-cp311-macosx_11_0_arm64.whl; mink-1.2.0-cp311-cp311-manylinux1_x86_64.manylinux_2_28_x86_64.manylinux_2_5_x86_64.whl; mink-1.2.0-cp311-cp311-manylinux2014_aarch64.manylinux_2_17_aarch64.manylinux_2_28_aarch64.whl; mink-1.2.0-cp311-cp311-musllinux_1_2_aarch64.whl; mink-1.2.0-cp311-cp311-musllinux_1_2_x86_64.whl; mink-1.2.0-cp311-cp311-win32.whl; mink-1.2.0-cp311-cp311-win_amd64.whl; mink-1.2.0-cp312-cp312-macosx_10_13_x86_64.whl; mink-1.2.0-cp312-cp312-macosx_11_0_arm64.whl; mink-1.2.0-cp312-cp312-manylinux1_x86_64.manylinux_2_28_x86_64.manylinux_2_5_x86_64.whl; mink-1.2.0-cp312-cp312-manylinux2014_aarch64.manylinux_2_17_aarch64.manylinux_2_28_aarch64.whl
Keywords: inverse, kinematics, mujoco
Tags
More Scientific/Engineering packages
NumPy provides an N-dimensional array object…
permissive · top 100 on PyPI
pandaspandas provides fast, flexible data structures…
permissive · top 100 on PyPI
scipyscipy provides numerical algorithms for…
permissive · top 100 on PyPI
scikit-learnscikit-learn provides a comprehensive Python…
permissive · top 1,000 on PyPI
dilldill extends Python's pickle module to…
permissive · top 1,000 on PyPI
multiprocessMultiprocess is an enhanced fork of Python's…
permissive · top 1,000 on PyPI
pin-pinkPink solves differential inverse kinematics for…
permissive · top 15,000 on PyPI
flexivrdkFlexiv Robotic Development Kit (RDK) provides…
permissive · top 15,000 on PyPI
robosuiterobosuite is a MuJoCo-powered simulation…
unclear · top 15,000 on PyPI
pinPython interface to a C++ rigid-body dynamics…
permissive · top 5,000 on PyPI
robot_descriptionsDownloads and caches open-source robot…
permissive · top 15,000 on PyPI
mujoco-warpGPU-accelerated physics simulation for robotics…
permissive · top 15,000 on PyPI
dex-retargetingTranslates human hand motion to robot hand…
permissive · top 15,000 on PyPI
libpinocchiolibpinocchio provides efficient rigid body…
permissive · top 15,000 on PyPI
mplibMPlib provides motion planning algorithms for…
permissive · top 15,000 on PyPI
toppraComputes time-optimal path parameterization for…
permissive · top 15,000 on PyPI