# kevinzakka/mink

Python inverse kinematics based on MuJoCo.

Repository: https://github.com/kevinzakka/mink
Canonical: https://ross.abutalabs.com/products/kevinzakka-mink
Homepage: https://kevinzakka.github.io/mink/
Language: Python
License: Apache-2.0
License Family: permissive
Topics: inverse-kinematics, mujoco, python
Last push: 2026-08-20T17:44:05+00:00

## Health v2 (maintenance only)
Score: 87/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 98, release rhythm 86, longevity 65
- inputs: {"age_days": 919, "days_push": 13, "days_rel": 16, "gap_med": 35.5, "n_releases_24m": 15}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1509, forks 116 (observed 2026-08-28T04:04:55.654312+00:00)

## What it is
mink is a Python library for differential inverse kinematics built on the MuJoCo physics engine. It solves task-space objectives as a quadratic program at each control step, respecting joint limits, velocity limits, collision avoidance, and loop-closure constraints.

## Use cases
- compute inverse kinematics for a robot arm in mujoco
- control a humanoid robot end effector to reach a target pose
- retarget human motion to a robot in real time
- avoid collisions while solving IK for a dual-arm setup
- teleoperate a robot with task-space objectives
- solve whole-body IK for a quadruped foot placement

## When to choose
- you already use MuJoCo for simulation and need IK at control rate
- you need composable task-space objectives with joint limits and collision avoidance
- you work with arms, humanoids, dexterous hands, or closed-chain mechanisms

## When to avoid
- you need global IK solutions rather than fast local differential IK
- your robot model is not in MuJoCo MJCF format
- you need a physics-agnostic IK solver independent of MuJoCo

## Facets
- artifact type: library
- maturity: active
- function: simulation, robotics, math
- domain: robotics, simulation
- platform: python, windows, cross-platform
- tags: inverse-kinematics, mujoco, quadratic-programming, motion-planning, real-time-control, robot-control, algorithms, linux, macos

## Member repositories
- kevinzakka/mink (main) score 87

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:55.654312+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T04:32:29.719827+00:00, confidence not recorded.
  - readme: https://github.com/kevinzakka/mink (fetched 2026-08-28T04:04:55.654312+00:00, sha 0d22d89f15cb)
  - homepage: https://kevinzakka.github.io/mink/ (fetched 2026-08-29T11:36:45.582736+00:00, sha 05aa789b3728)
  - registry_pypi: https://pypi.org/pypi/mink/json (fetched 2026-08-29T11:36:45.591778+00:00, sha b0c9eb3f90cd)
- Data as of 2026-08-30T08:39:29.467469+00:00.