Domain adaptation framework for visual RL using adaptive imagination for scarce data.

A dataset of image-question pairs designed to improve visually faithful reasoning in VLMs.

A multimodal foundation model family purpose-built for adversarial content and AI safety.

RL post-training framework for multi-turn image editing in unified multimodal models.

A graphical desktop OS that fits on a 1.44MB floppy disk.

A modular image-to-3D generation system for interior design assets.

Provide coding agents with the decisions your team made.

A local multi-agent harness for dispatching tasks and orchestrating local LLM agents.

A lightweight server monitoring tool.

A scheme for protecting dense embeddings during semantic search using geometric obfuscation and homomorphic encryption.

A video diffusion transformer for multispectral Earth observation forecasting incorporating physical conditioning.

A domain security scanner that checks for exposed secrets, TLS issues and DNS/email auth vulnerabilities.

A tool for deployment and infrastructure management.

A vanilla-JS agent UI library with native WebMCP support.

A true mobile Linux distribution for smartphones and smartphones with older hardware.

Multi-agent framework for adaptive personalized AI-assisted learning and embodied teaching.

Hybrid neuro-symbolic gating for irony detection in social media texts.

A framework for generating visually grounded, multi-shot short dramas from plots.

Retrieval-constrained LLM adjudication for clinical value set authoring.

An internal platform for visualizing the distillation pipeline.