Dev Tools
VelesDB
Explainable, local-first memory engine designed for AI agents.
★ 97 starsNOASSERTION
View guide →A GPU-accelerated, differentiable physics engine on NVIDIA Warp for robotics simulation.
newton is an open-source dev tools project by newton-physics (newton-physics/newton on GitHub). It has 5,736 GitHub stars and 715 forks, is released under the Apache-2.0 license, and was last updated on GitHub on 8 October 2026.
Decide what you are simulating — a robot, a mechanism, a scene — and whether you need GPU acceleration or differentiability.
Install the package with the examples extra from PyPI.
Confirm your NVIDIA GPU and driver meet the Warp requirements, or plan to run on CPU.
Launch the examples browser, pick a scene such as the URDF or pendulum example, and view it with the GL or USD viewer.
Physics runs on the GPU via NVIDIA Warp for scalable, rapid-iteration robotics simulation.
Integrates MuJoCo Warp as its primary simulation backend, generalizing Warp's deprecated warp.sim module.
Designed for differentiability, supporting gradient-based optimization and learning workflows.
Works with OpenUSD scenes and can save simulation output to USD files through the USD viewer.
User-defined extensibility lets researchers add their own solvers, forces and components.
Ships basic, robot, cloth, cable and viewer examples runnable through a single python -m newton.examples entry point.
Clone Newton if you want the source; most users install from PyPI in the next step.
git clone https://github.com/newton-physics/newton.git
cd newtonInstall Newton and its example dependencies.
pip install "newton[examples]"List the examples and run one on your chosen device with the USD viewer.
python -m newton.examples --list
python -m newton.examples basic_urdf --device cuda:0Newton is a community-built Linux Foundation project initiated by Disney Research, Google DeepMind and NVIDIA.
Yes — the code is Apache-2.0 licensed (documentation is CC-BY-4.0).
No — it is a local simulation library. You need suitable hardware: an NVIDIA Turing+ GPU for CUDA acceleration, or you can run on CPU.
Newton builds on Warp with MuJoCo Warp as its primary backend, emphasizing GPU-native, differentiable, extensible simulation in Python rather than a full commercial simulator application.
pip install "newton[examples]" and run python -m newton.examples to browse and launch examples.
Explainable, local-first memory engine designed for AI agents.
PostgresML — machine learning inside PostgreSQL with GPU acceleration: train, embed and run LLMs in SQL.
NVIDIA NeMo Data Designer: build validated, multimodal synthetic datasets from scratch or from seed data.
An AI compute engine: a distributed runtime plus libraries for scaling ML and LLM workloads
MCP graph engine that compounds context across AI coding assistants.
Open-source Rust agent engine and terminal client with provider choice, approvals and receipts
The most-starred other Dev Tools projects in the LibreAI catalogue — the realistic shortlist if newton is not the right fit.
The open source coding agent
The model-definition framework behind modern AI — thousands of pretrained transformers for text, vision, and audio.
Anthropic's agentic coding tool that lives in your terminal and works across your whole codebase
Cross-platform desktop all-in-one manager for Claude Code, Codex, OpenCode and other coding agents
OpenAI's lightweight coding agent that runs locally in your terminal
Free and open — grab the source, star the repo, and start creating.