# dartsim/dart

Research-focused C++23 physics engine for robotics, animation, and machine learning, with Python bindings

Repository: https://github.com/dartsim/dart
Canonical: https://ross.abutalabs.com/products/dartsim-dart
Homepage: https://dart.readthedocs.io
Language: C++
License: BSD-2-Clause
License Family: permissive
Topics: robotics, computer-animation, simulation, kinematics, dynamics, lie-group, inverse-kinematics, urdf, collision-detection, gazebo, mjcf, physics-engine, python-bindings, sdf, multibody, nanobind, rigidbody, cpp23
Last push: 2026-09-03T01:52:19+00:00

## Health v2 (maintenance only)
Score: 98/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 100, release rhythm 93, longevity 100
- inputs: {"age_days": 5462, "days_push": 0, "days_rel": 45, "gap_med": 8.5, "n_releases_24m": 19}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1199, forks 304 (observed 2026-09-03T02:15:11.107839+00:00)

## What it is
DART (Dynamic Animation and Robotics Toolkit) is an open-source, research-focused C++ physics engine for robotics, animation, and machine learning, with Python bindings via dartpy. It provides transparent kinematics, dynamics, collision detection, and constraint solving using generalized coordinates and Featherstone's Articulated Body Algorithm.

## Use cases
- simulate articulated rigid body dynamics for robotics research
- compute inverse kinematics for robot arms and humanoids
- load and simulate URDF, SDF, or MJCF robot models
- physics-based character animation with stable dynamics
- build custom simulators for machine learning and reinforcement learning
- collision detection for multibody systems

## When to choose
- you need transparent, research-grade dynamics rather than a black-box simulator
- you work with articulated rigid body systems and want direct access to dynamics quantities
- you need accurate, stable motion via Featherstone's algorithm
- you want C++ performance with Python bindings for ML workflows

## When to avoid
- you need a production-ready stable release on the main branch - use DART 6 LTS instead
- you want a turnkey game physics engine with broad asset tooling
- you need GPU-accelerated massively parallel simulation
- your project depends on a different ecosystem like MuJoCo or PyBullet conventions

## Facets
- artifact type: library
- maturity: active
- function: simulation, computer-vision, robotics, graphics
- domain: robotics, simulation, machine-learning, graphics
- platform: cpp, python, windows, cross-platform
- tags: physics-engine, kinematics, dynamics, inverse-kinematics, collision-detection, urdf, mjcf, sdf, multibody-dynamics, featherstone, lie-groups, python-bindings, nanobind, computer-animation, articulated-rigid-bodies, game-development, linux, macos

## Member repositories
- dartsim/dart (main) score 98

## Provenance
- Observed fields: from GitHub, fetched 2026-09-03T02:15:11.107839+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 2026-08-30T06:22:02.029796+00:00, confidence not recorded.
  - readme: https://github.com/dartsim/dart (fetched 2026-09-03T02:15:11.107839+00:00, sha 9db41bd6000a)
- Data as of 2026-08-30T08:39:29.467469+00:00.