RobotLocomotion/drake

Health v2 · maintenance only

Flags: no_license

Adoption not part of the score

4158 stars · 1382 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

Drake is a C++ toolbox (with Python bindings) for model-based design and verification of robotics systems, developed by the Robot Locomotion Group at MIT and now led by the Toyota Research Institute. It provides tools for analyzing robot dynamics, simulating complex physical phenomena like friction and contact, and building optimization-based planning, control, and analysis algorithms.

Use cases

• simulate robot dynamics with contact and friction
• design and verify control systems for robots
• solve trajectory optimization problems
• compute multibody kinematics and dynamics
• prototype robotics algorithms in Python
• model and analyze dynamical systems
• solve mathematical programs for planning and control

When to choose

• you need rigorous, optimization-based robot dynamics simulation and analysis rather than a black-box simulator
• you are doing research or advanced development in robot planning, control, or verification
• you want analytical gradients, sparsity structure, and uncertainty quantification exposed from the governing equations
• you need a well-supported open-source toolbox with C++ performance and Python prototyping

When to avoid

• you need a lightweight, easy-to-learn game-style physics engine for simple simulations
• you require Windows or other unsupported platforms in official configurations
• you want a quick plug-and-play robotics middleware rather than a deep modeling and analysis toolkit
• your team cannot work with a large C++ codebase and its build requirements

Facets

library · maturity stable

simulation robotics machine-learning math robotics simulation cpp python model-based-design control-systems multibody-dynamics verification trajectory-optimization c-plus-plus python-bindings optimization algorithms research linux macos

Member repositories

For agents

markdown · JSON · MCP: product_card(name="RobotLocomotion/drake")