World Models. Embodied AI.
Tiny JEPA world models that think live
Tiny JEPA world models that think live
PI0.5 manipulation results on all four LIBERO suites
Explore robot demonstration episodes with video and pose details
Explore robotic assembly tasks with videos and instructions
Generate robot actions from natural language instructions
A web-based robotics control
Diffusion VLA predicting robot actions and future frames
Explore and download open VLN datasets and robot tools
Generate AI-powered YouTube Shorts videos
Race a 3D car through a sunset track
Multilayer-geometry 3D from a single image or 16-frame clip
MolmoPoint - Image & Video Pointing & Tracking
Run Maple-Preview privately in your browser with WebGPU