Physical Ai Python Packages

ovstage

A C and Python library providing a shared, high-performance, vectorized, GPU-capable scene data substrate for OpenUSD data across Omniverse Libraries (physics, rendering, sensors, animation).

ovrtx

A C and Python library for physically accurate, real-time, sensor simulation and visualization using NVIDIA Omniverse RTX

roboto

Data Analytics for Robotics and Physical AI

hflow

SDK for robotics teams to verify the quality of their data used for AI model training.

inspect-robots

Open source evals for physical AI. Run any LLM/VLA on any arm/humanoid against any real/sim benchmark.

quackd

One CLI for all your robots. Connect them, command them, and let them work together, each with an LLM for a brain (Claude, OpenAI, Gemini, Grok, or local via Ollama or vLLM), VLAs for arms and decision LLMs (Jev, Laya, Kev). Drives Microduck, Open Duck Mini, LeRobot, XLeRobot, AlohaMini, ToddlerBot, or any ROS 2 base from a laptop, offboard.

robot-reel

Review Physical AI with original evidence: compare VLM explanations against recorded robot outcomes, replay SmolVLA failures, and inspect Newton/Genesis simulations offline.

quackd-lerobot

One CLI for all your robots. Connect them, command them, and let them work together, each with an LLM for a brain (Claude, OpenAI, Gemini, Grok, or local via Ollama or vLLM), VLAs for arms and decision LLMs (Jev, Laya, Kev). Drives Microduck, Open Duck Mini, LeRobot, XLeRobot, AlohaMini, ToddlerBot, or any ROS 2 base from a laptop, offboard.

quackd-xlerobot

One CLI for all your robots. Connect them, command them, and let them work together, each with an LLM for a brain (Claude, OpenAI, Gemini, Grok, or local via Ollama or vLLM), VLAs for arms and decision LLMs (Jev, Laya, Kev). Drives Microduck, Open Duck Mini, LeRobot, XLeRobot, AlohaMini, ToddlerBot, or any ROS 2 base from a laptop, offboard.

quackd-toddlerbot

One CLI for all your robots. Connect them, command them, and let them work together, each with an LLM for a brain (Claude, OpenAI, Gemini, Grok, or local via Ollama or vLLM), VLAs for arms and decision LLMs (Jev, Laya, Kev). Drives Microduck, Open Duck Mini, LeRobot, XLeRobot, AlohaMini, ToddlerBot, or any ROS 2 base from a laptop, offboard.

fullseye

Fullseye — a numpy-native operator library for vision, 3-D and measurement: 1,500+ typed ops spanning 2-D imaging, point clouds/meshes/volumes, optics, tomography, interferometry, acoustics, colour and metrology, plus an evolutionary pipeline designer with honest holdout gating. Ships as an AI/RAG operator corpus.

quackd-microduck

One CLI for all your robots. Connect them, command them, and let them work together, each with an LLM for a brain (Claude, OpenAI, Gemini, Grok, or local via Ollama or vLLM), VLAs for arms and decision LLMs (Jev, Laya, Kev). Drives Microduck, Open Duck Mini, LeRobot, XLeRobot, AlohaMini, ToddlerBot, or any ROS 2 base from a laptop, offboard.

feral-ai

one Memory across your devices and apps

quackd-alohamini

One CLI for all your robots. Connect them, command them, and let them work together, each with an LLM for a brain (Claude, OpenAI, Gemini, Grok, or local via Ollama or vLLM), VLAs for arms and decision LLMs (Jev, Laya, Kev). Drives Microduck, Open Duck Mini, LeRobot, XLeRobot, AlohaMini, ToddlerBot, or any ROS 2 base from a laptop, offboard.

quackd-rosbridge

One CLI for all your robots. Connect them, command them, and let them work together, each with an LLM for a brain (Claude, OpenAI, Gemini, Grok, or local via Ollama or vLLM), VLAs for arms and decision LLMs (Jev, Laya, Kev). Drives Microduck, Open Duck Mini, LeRobot, XLeRobot, AlohaMini, ToddlerBot, or any ROS 2 base from a laptop, offboard.

quackd-open-duck

One CLI for all your robots. Connect them, command them, and let them work together, each with an LLM for a brain (Claude, OpenAI, Gemini, Grok, or local via Ollama or vLLM), VLAs for arms and decision LLMs (Jev, Laya, Kev). Drives Microduck, Open Duck Mini, LeRobot, XLeRobot, AlohaMini, ToddlerBot, or any ROS 2 base from a laptop, offboard.

safety-harness

A perception-grounded safety harness for physical AI actuators: default-deny precondition gate, permission not detection.

ferrotherm

Thermodynamic computing in pure Rust: energy-based models, block-Gibbs, parallel tempering, thermodynamic linear algebra, stochastic differentiable programs, a variational compiler onto device topologies, and a joules ledger. Zero deps, wasm-clean, verified against exact physics — and it reaches real Ising silicon through the same trait as a CPU.

ml-dash

Scalable Training Telemetry and Metrics Visualization

inspect-robots-agent

Open source evals for physical AI. Run any LLM/VLA on any arm/humanoid against any real/sim benchmark.

tatolab-stream

Control runtime for building things in the Physical AI space. Everything from self piloting vehicles to humanoid robots. Purposefully built to fuse LLM / AI control policies with your code

forge-robotics

Robotics Data Toolkit | Convert between robotics dataset formats (RLDS, LeRobot v2/v3, Zarr, HDF5, Rosbag). Inspect, visualize, and analyze datasets. Works with HuggingFace Hub. Built for OpenVLA, Octo, LeRobot, and Diffusion Policy workflows.

tatolab-webrtc

Control runtime for building things in the Physical AI space. Everything from self piloting vehicles to humanoid robots. Purposefully built to fuse LLM / AI control policies with your code

otwin

Otwin is an open-source physics engine for Engineering sytems