Nanosaur2-670M
well9472 · Kit
Makes illustrations from text prompts with a small 670M-parameter diffusion model, run through ComfyUI. Built as a low-budget research model for people who wan…
New
Overview
Simulates the physics of jointed bodies that touch their surroundings, fast and accurately, for work in robotics, biomechanics, animation and machine learning. Comes with a C API, Python bindings and a Unity plug-in.
mujoco is an open-source project on GitHub in the Developer category, listed by Mindket. It is free. This is version 1, last updated on . Install it from GitHub as its README says. Get Mindket’s setup notes on your Mindket account, as a zip.
At a glance
- You need
- Git, to get the code
- A C or Python project that will call the library, or a Unity project for the plug-in
- Runs in
- C programs: the core library has a C API
- Python: the project provides bindings
- Unity: the project provides a plug-in for the game engine
- Desktop windows and web browsers: interactive visualization runs in both
- Start with
git clone https://github.com/google-deepmind/mujoco.git cd mujoco
Platform
Desktop
About
What it is
MuJoCo, short for Multi-Joint dynamics with Contact, is a physics engine for simulating jointed bodies as they touch and move through their surroundings. It aims for speed and accuracy. Google DeepMind maintains it, and it is written in C++.
Who it suits
Researchers and developers who need fast, accurate simulation of jointed structures. The project names machine learning, robotics, biomechanics, animation and graphics as its areas.
| Industry | Who uses it | How they use it | Example |
|---|---|---|---|
| Robotics | Robotics researchers | Simulate robot bodies in contact with their surroundings | A robot arm pushing a block |
| Biomechanics | Biomechanics researchers | Model a jointed body moving through its environment | A simulated limb under load |
| Animation and graphics | Animators and graphics developers | Drive the motion of jointed figures with physics | A character figure falling to the floor |
| Machine learning | Machine learning engineers | Run fast simulations as a test bed for learning experiments | A simulated body used for repeated trials |
You can skip it if you do not write code, since it is built for researchers and developers.
Where to use it
- C programs: the core library has a C API
- Python: the project provides bindings
- Unity: the project provides a plug-in for the game engine
- Desktop windows and web browsers: interactive visualization runs in both
- Your browser: the project offers a way to try it online
| Use case | Who | What they get back |
|---|---|---|
| Robot simulation | Robotics researchers | Contact-aware motion of jointed bodies |
| Interactive viewing | Developers | Live visualization in a desktop window or a browser |
| Physics calculations | Researchers | Utility functions that compute physics-related quantities |
How to set it up
- Git, to get the code
- A C or Python project that will call the library, or a Unity project for the plug-in
- Clone the repository with git.
- Follow the setup in its README.
- Read the documentation it points to for the details of the API.
Run these in a terminal:
# Copy the project to your computer git clone https://github.com/google-deepmind/mujoco.git # Go into its folder cd mujoco
How to use it
You work through the C API, the Python bindings or the Unity plug-in. A built-in XML compiler prepares the low-level data structures the runtime simulation operates on. Utility functions then compute physics quantities from the results, and the viewer shows the motion live.
For newcomers and experts
A physics engine works out how connected parts move and push on each other. A good first try is the project's browser demo, which lets you watch a simulation before you install anything.
If you work in simulation, the runtime module is tuned for speed and operates on low-level, preallocated data structures. The cost of that tuning is a low-level interface, so expect to work close to the data.
- Describe the model for the XML compiler
- The compiler preallocates low-level data structures
- The runtime module simulates the physics
- View it live or compute physics quantities
Good to know
- License: Apache-2.0
- Cost to run: none named
- Maintainer: Google DeepMind
- Audience: researchers and developers, through a C API
- What the kit holds: notes on the project and where its code is. Nothing in it runs by itself.
README and files
mujoco
Simulates the physics of jointed bodies that touch their surroundings, fast and accurately, for work in robotics, biomechanics, animation and machine learning. Comes with a C API, Python bindings and a Unity plug-in.
Made by google-deepmind on GitHub. These notes are Mindket's, written from its README; the project's own files and license are in its repository.
About
What it is: MuJoCo, short for Multi-Joint dynamics with Contact, is a physics engine for simulating jointed bodies as they touch and move through their surroundings. It aims for speed and accuracy. Google DeepMind maintains it, and it is written in C++.
Who it suits: Researchers and developers who need fast, accurate simulation of jointed structures. The project names machine learning, robotics, biomechanics, animation and graphics as its areas.
Industry | Who uses it | How they use it | Example
Robotics | Robotics researchers | Simulate robot bodies in contact with their surroundings | A robot arm pushing a block
Biomechanics | Biomechanics researchers | Model a jointed body moving through its environment | A simulated limb under load
Animation and graphics | Animators and graphics developers | Drive the motion of jointed figures with physics | A character figure falling to the floor
Machine learning | Machine learning engineers | Run fast simulations as a test bed for learning experiments | A simulated body used for repeated trialsYou can skip it if you do not write code, since it is built for researchers and developers.
Where to use it:
- C programs: the core library has a C API
- Python: the project provides bindings
- Unity: the project provides a plug-in for the game engine
- Desktop windows and web browsers: interactive visualization runs in both
- Your browser: the project offers a way to try it online
Use case | Who | What they get back
Robot simulation | Robotics researchers | Contact-aware motion of jointed bodies
Interactive viewing | Developers | Live visualization in a desktop window or a browser
Physics calculations | Researchers | Utility functions that compute physics-related quantitiesHow to set it up:
- Git, to get the code
- A C or Python project that will call the library, or a Unity project for the plug-in
- Clone the repository with git.
- Follow the setup in its README.
- Read the documentation it points to for the details of the API.
How to use it: You work through the C API, the Python bindings or the Unity plug-in. A built-in XML compiler prepares the low-level data structures the runtime simulation operates on. Utility functions then compute physics quantities from the results, and the viewer shows the motion live.
For newcomers and experts: A physics engine works out how connected parts move and push on each other. A good first try is the project's browser demo, which lets you watch a simulation before you install anything.
If you work in simulation, the runtime module is tuned for speed and operates on low-level, preallocated data structures. The cost of that tuning is a low-level interface, so expect to work close to the data.
Describe the model for the XML compiler
The compiler preallocates low-level data structures
The runtime module simulates the physics
View it live or compute physics quantitiesGood to know:
- License: Apache-2.0
- Cost to run: none named
- Maintainer: Google DeepMind
- Audience: researchers and developers, through a C API
- What the kit holds: notes on the project and where its code is. Nothing in it runs by itself.
On GitHub, 9 November 2026
- Stars: 16,450 (41 that day)
- Forks: 1,953
- Watching: 138
- Open issues and pull requests: 308
- Language: C++
- License: Apache-2.0
- Topics: mujoco, physics, robotics
- Trending: #8 repository of the day on GitHub trending (9 November 2026)
Links
- Code: https://github.com/google-deepmind/mujoco
- Website: https://mujoco.org/
- Releases: https://github.com/google-deepmind/mujoco/releases
- Issues: https://github.com/google-deepmind/mujoco/issues
- Star history: https://www.star-history.com/#google-deepmind/mujoco&Date
Intro videos
- MuJoCo simulate tutorial: https://www.youtube.com/watch?v=P83tKA1iz2Y
Get the code
git clone https://github.com/google-deepmind/mujoco.gitThen follow the setup in its README. Nothing in this kit runs by itself.
README.md 4 KB
What’s new
Reviews
What people say
Posts and comments that name this project on Hacker News, Bluesky, X, Reddit, Facebook and YouTube, found by a weekly search and on Trendshift. Mindket doesn’t write or check them. Open one to read it where it was posted.
prometheus1992MicroduckNot sure many people know this - most of the robotics stuff that you see in the news is based on MujoCo (Multi-Joint dynamics with Contact) engine, that is used primarily to create simulated environments in which these robots learn the RL policies. This project is maintained by Google deepmind -…GitHub Trending📈 Today's GitHub Trending: 1. https://github.com/anthropics/skills - Python, 1052 stars today 2. https://github.com/meta/astryx - TypeScript, 3324 stars today 3. https://github.com/sherlock-project/sherlock - Python, 4294 stars today 4. https://github.com/google-deepmind/mujoco - C++, 579 stars […]1 likewesturnerElectronic nose for indoor mold detection and identificationMany of the described robots in the mujoco_menagerie gallery could probably do the job: https://github.com/google-deepmind/mujoco_menagerie The sensor recovery time limits the sampling time. A 28 second sensor recovery time with SnO2-Gr (graphene oxide) would be more useful for sampling larger…デイリーGitHubトレンド今日のGitHubトレンド google-deepmind/mujoco MuJoCoは、ロボティクスや機械学習などの研究開発に向けた汎用物理エンジンです。 関節構造と環境の相互作用を高速かつ正確にシミュレートすることを目的としています。 C APIによる高性能な動作に加え、可視化機能やPythonバインディング、Unityプラグインなどを提供しています。westurnerRobots – Comprehensive catalog of 115 humanoid and quadruped robotsmujoco_menagerie includes Mujoco physics simulator MJCF models for 54 robot models; 9 humanoids, 8 quadrepeds: https://github.com/google-deepmind/mujoco_menagerie#menageri... mujoco_menagerie can also load any of the 135+ models supported by robot_descriptions FWIU pypi:robot_descriptions:…GitHub Trending📦 google-deepmind / mujoco ⭐ 16,148 (+26) 🗒 C++ Multi-Joint dynamics with Contact. A general purpose physics simulator.xnxWhy today's humanoids won't learn dexterityGoogle has done training in simulation: https://x.company/projects/everyday-robots/#:~:text=other%20... I believe this is the most popular tool now: https://github.com/google-deepmind/mujocoGlonceGoogle DeepMind unveils Gemini 4 Argon, a physical reasoning model for robotics DeepMind says Gemini 4 Argon scores 87% on the ERQA embodied reasoning benchmark, ahead of every other frontier model, and shows example prompts for controlling a Spot robot, building MuJoCo digital twins and…westurnerNitro: A tiny but flexible init system and process supervisorFor workstations with GPUs and various kernel modules, rpm-ostree + GRUB + Native Containers for the rootfs and /usr and flatpaks etc on a different partition works well enough. ostree+grub could be much better at handling failover like switches and rovers that then need disk space for at least…اسلیم کلیک هاWell that's a good sign. Had an idea for a better way to approach this wasm kernel business (vaguely: arena hub and spoke) and asked why I haven't read about anything else that works like this and the only result was github.com/google-deepm... (opencvjs doesn't count, pictures is ez)westurnerGemini Robotics On-Device brings AI to local robotic devicesmujoco_menagerie has Mujoco MJCF XML models of various robots. google-deepmind/mujoco_menagerie: https://github.com/google-deepmind/mujoco_menagerie mujoco_menagerie/aloha: https://github.com/google-deepmind/mujoco_menagerie/tree/mai...Hacker News Top StoriesMuJoCo – Advanced Physics Simulation | Discussion
More from Mindket
See all from Mindketwaqf-ocr-hand-written-v1
Waqf-AI · Kit
Reads handwritten Arabic text from images and returns it as text. A fine-tuned PaddleOCR-VL model for people digitizing Arabic handwriting.
New
Wan2.2-TI2V-5B-Diffusers
Wan-AI · Kit
Makes video clips from a text prompt or an image plus text, at up to 720P and 24 frames per second, in the Diffusers format for a single consumer GPU.
New
Similar in Developer
See all in DeveloperWan2.2-S2V-14B
Wan-AI · Kit
Generates cinematic video of a character driven by an audio track, and can also lip-sync existing video. For video creators, animators and researchers with a l…
New
Wan2.2-Animate-14B
Wan-AI · Kit
Animates a character or swaps one into a video by copying a performer's movement and expressions, as one model from the Wan2.2 video family. Use only faces and…
New
d3-turbo
vllm-sr · Kit
Answers choice, yes-or-no, score, box and tracking questions about text, images or video in one call, with a probability for each answer. For people building r…
New