Stars
HaWoR: World-Space Hand Motion Reconstruction from Egocentric Videos
[CVPR 2026] UniDex: A Robot Foundation Suite for Universal Dexterous Hand Control from Egocentric Human Videos
Code for NeurIPS 2022 Datasets and Benchmarks paper - EgoTaskQA: Understanding Human Tasks in Egocentric Videos.
Interactive World Simulator for Robot Policy Training and Evaluation
GigaWorld-0: World Models as Data Engine to Empower Embodied AI
Your own personal AI assistant. Any OS. Any Platform. The lobster way. 🦞
RynnBrain: Open Embodied Foundation Models
RynnScale: Scalable VLM and VLA Development Kits
Fast, Sharp & Reliable Agentic Intelligence
[ICLR 2026] The offical Implementation of "Soft-Prompted Transformer as Scalable Cross-Embodiment Vision-Language-Action Model"
Step3-VL-10B: A compact yet frontier multimodal model achieving SOTA performance at the 10B scale, matching open-source models 10-20x its size.
MAI-UI: Real-World Centric Foundation GUI Agents ranging from 2B to 235B
Official code of Motus: A Unified Latent Action World Model
DeepEP: an efficient expert-parallel communication library
TurboDiffusion: 100–200× Acceleration for Video Diffusion Models
MMSI-Video-Bench: A Holistic Benchmark for Video-Based Spatial Intelligence
Octo is a transformer-based robot policy trained on a diverse mix of 800k robot trajectories.
Miles is an enterprise-facing reinforcement learning framework for LLM and VLM post-training, forked from and co-evolving with slime.
A curated collection of fun and creative examples generated with Nano Banana & Nano Banana Pro🍌, Gemini-2.5-flash-image based model. We also release Nano-consistent-150K openly to support the commu…
[ICLR 2026] MMSI-Bench: A Benchmark for Multi-Image Spatial Intelligence
Code for [AAAI 2026] AffordDex: Towards Affordance-Aware Robotic Dexterous Grasping with Human-like Priors
A framework aiming to bridge fast robot prototyping, predefined motion primitives, heterogeneous teleoperation, data collection, and flexible deployment across diverse robot platforms.