AI tools for Physics

FlowHMR reconstructs 3D human motion from monocular video. It targets motion that can be tracked in a physics simulation and provides tools to preview the recovered sequence and

RoboSkills is an open collection of lightweight agent workflows and tools for robotics. It packages five working approaches: automated research, agent-as-policy control, agentic

Viggle JST-2 is a physics-native world foundation model for generating scenes, characters, and their motion. It represents persistent matter as 3D atoms with explicit positions

FlashDexRetarget converts collections of human hand-object demonstrations into physically grounded robot-hand trajectories. One reference-conditioned reinforcement-learning poli

OVOW, or One Video One World, reconstructs a monocular video as a structured 4D scene with separate mesh objects and motion. It targets simulation-ready output rather than only

StreamForce is a streaming video generation framework that lets users apply and modify physical forces while a video is being generated. Starting from a single image, it support

TriSplat is a feed-forward 3D scene reconstruction method designed for producing simulation-ready mesh-like scene reconstructions from sparse views. It focuses on making this wo

PhysX-Omni is a unified simulation-ready physical 3D generation framework designed for generating rigid, deformable, and articulated 3D assets with physical properties. It focus

CubePart is an open-vocabulary part-controllable 3D generator designed for creating multi-part 3D mesh objects from text prompts and part schemas. It focuses on making this work

PhysForge is a framework for generating physics-grounded 3D assets for interactive virtual worlds and embodied AI environments. It addresses a limitation of many 3D generation s

PhyMotion is a structured 3D motion reward for improving and evaluating physics-grounded human video generation. It targets a major weakness in video generation: models can prod

OpenGame is an open-source agentic coding framework built to turn natural-language game ideas into playable web games. It is designed for the hard part of game creation: coordin

MMPhysVideo is a research system focused on multimodal physical reasoning for video. It is designed for tasks where video models must understand not only appearance and motion,

TutorGPT functions as a powerful, always-available personal tutor, meticulously engineered to assist students in overcoming academic hurdles across a broad spectrum of subjects.

MOVA is a digital human model designed to simulate realistic human-like movements and interactions. It is capable of generating highly realistic animations and can be used in va

VIGA is a multimodal agent that treats vision as inverse graphics, reconstructing images into editable 3D scene programs via an analysis-by-synthesis loop in Blender.

This advanced platform revolutionizes content creation by effortlessly transforming simple text prompts and static images into high-quality, professional videos. Leveraging cutt

UltraShape 1.0 is a high-fidelity 3D shape generation system built around a scalable two-stage diffusion framework designed specifically for high-quality geometry synthesis. It

Stable Video Infinity (SVI) represents a breakthrough in video generation technology, enabling the creation of videos of any length while maintaining exceptional temporal consis

ConceptViz is a specialized visual creation tool engineered to bridge the gap between complex scientific concepts and clear, digestible classroom materials. Designed for educato