Intelligence as Relationship
The official open-source organization for Chinoba — an open research ecosystem building the infrastructure for Human–AI Coordination.
Chinoba is built on a single idea: intelligence is not a property of a model — it emerges in relationship. It lives in how AI, humans, organizations, and society trust each other, share what they know, decide, and act together.
So the work here is not better models. It is the infrastructure of the relationship — trust, knowledge, accountable decisions, and the coordination that lets humans and AI act together in the open.
AI is not prediction. It is decision.
Explore Chinoba's complete Lines of Inquiry.
➡ Browse the complete Research Hub
chinoba-lab hosts the official, open-source implementation of Chinoba,
organized into five layers around a single principle: the runtime is the only
decision authority.
| Layer | What it does |
|---|---|
| Interaction | Turns messages and evidence into structured signals |
| Core / Decision Runtime | The single point of decision — evaluates, bounds, records |
| Execution | Enacts finalized decisions across agents, tools, and systems |
| Learning | Learns from ledgered traces and emits better signals |
| Design | Designs, simulates, and improves decision systems |
Every layer produces signals, follows decisions, or makes them understandable. Only the runtime decides.
→ The full architecture and decision flow lives at chinoba.org.
Chinoba's repositories are organized into reusable building blocks — each a self-contained piece of the decision infrastructure, meant to be composed rather than reinvented.
| Category | Description | Status |
|---|---|---|
| Starter Kits | End-to-end templates for the signal → decision → trace flow | Planned |
| Decision Runtime | The Decision OS kernel — the single point of decision | Planned |
| Schemas | Shared contracts for signals, decisions, and traces | Planned |
| Examples | Worked scenarios showing the runtime in practice | Planned |
| Tools | Utilities for building, inspecting, and testing decisions | Planned |
| Infrastructure | Deployment, ledgering, and coordination substrate | Planned |
| SDKs | Language bindings for connecting systems to the runtime | Planned |
Research
↓
Research Hub
↓
Open Source
↓
Decision Runtime
↓
Applications
Ideas originate in research, are documented and made navigable in the Research Hub, mature into reusable open-source components, converge on the Decision Runtime as the single decision authority, and finally power real-world AI applications.
- Starter Kit (runtime-connected) — the full signal → decision → trace flow.
- Starter Kit (standalone) — local, self-contained decision logic.
- decision-runtime-core — the Decision OS kernel, the center of the system.
Browse the repositories below, each tagged with its layer (layer-core,
layer-interaction, …) and stream (stream-decision, stream-coordination).
Chinoba began — and continues to experiment — on the founder's personal account,
Masao-Watanabe-AI. That account is the
Bench: experiments, prototypes, and the archive of earlier work. This
organization is the Commons: the official, supported ecosystem. Ideas
graduate from one to the other.
- Website — chinoba.org
- Research Blog — deus-ex-machina-ism.com
- Library (JP & EN) — chinoba.org/library
- YouTube — Chinoba Research Archive
- LinkedIn — masao-watanabe-ai
Chinoba · Intelligence as Relationship · chinoba-lab