RoboClaw is an Embodied AI Assistant.

We release Evo-RL, the opensource real-world offline RL on So-101 and AgileX PiPER for easier reproduction.

ConLA: Contrastive Latent Action Learning from Human Videos for Robotic Manipulation

The official implementation of VLA-Pruner: Temporal-Aware Dual-Level Visual Token Pruning for Efficient Vision-Language-Action Inference.

This website is for the collection of VLA SOTA results.

Evo-1: Lightweight Vision-Language-Action Model with Preserved Semantic Alignment

U-Arm: Lerobot-Everything-Cross-Embodiment-Teleoperation

Evo-0: Vision-Language-Action Model with Implicit Spatial Understanding.

The open-source CapCut alternative

🔥🔥First-ever hour scale video understanding models

RoboFAC: A Comprehensive Framework for Robotic Failure Analysis and Correction

STI-Bench : Are MLLMs Ready for Precise Spatial-Temporal World Understanding?

🔥🔥MLVU: Multi-task Long Video Understanding Benchmark