rUv

Governed Evolutionary Software — Autogenous turns runtime failures into verified, portable, reversible software adaptations. AGL typed mutations, antibody packages, hard-gate promotion, automatic rollback.

Invisible, key-verifiable watermarks for AI/LLM-generated text — SynthID-Text-style, in Rust + WebAssembly (Node + browser). Live playground included.

Latent communication fabric for continuously evolving agent collectives — training-free orthogonal alignment of hidden states as a network primitive.

ruvnet-only control plane for running an AI-agent company end-to-end — org-chart, budget-gated heartbeats, approval gates, signed audit trail. Built on ruflo, metaharness, ruvector, and agentbbs.

Reformat a live LLM token stream inflight — separate thinking from answer, collapse repetition, normalize. Google Vertex / OpenRouter / meta-llm. Rust + WASM, built on ruvnet/midstream.

OpenAVO is an open harness for long horizon autonomous reasoning, combining prediction, memory, supervision, verification, and adaptive action loops to improve agent reliability on ARC AGI style interactive tasks and other complex environments.

A MoE expert separation workbench for building, testing, and routing smaller specialist models with MetaHarness, Ruflo, and LatentMesh state exchange. Includes reproducible benchmarks, quality gates, architecture decisions, and animated explainers.

ruOS — an agentic desktop (a desktop that runs itself): computer-use MCP server + ruvnet-stack skills.

A friendly animated robot pet for the ChatGPT and Codex desktop pet system.

Self-optimizing Raspberry Pi 5 / CrowPi sensor box in Rust. One trait over every device with verification provenance in the type system, plus a governed evolution loop (autogenous fitness gate, canary, verified rollback, signed witness chain).

Group Field Theory proposes quantum building blocks that do not live in a preexisting universe. Under the right collective conditions, their relations become the universe.

APx, or Agentic Power, is a framework for measuring the useful work produced by agent systems—not just how long agents run or how many tasks they complete. It evaluates output by task, quality, cost, safety, and the amount of human effort required.

Independent, privacy-governed multivariate time-series forecasting in Rust — Burn-based CPU/CUDA/WGPU model, cryptographically-verified hyperparameter promotion, honest accuracy research in progress

Open source macOS Bluetooth control, animated displays, and RuView sensing examples for the Divoom MiniToo

An interactive dot matrix black hole simulation and real exoplanet atlas built with React, Three.js, Vinext, and Cloudflare Workers.

A perpetual agentic runtime for Artificial General Intelligence built on the ruvstack