Build systems that can see the world clearly, reason about change, and act with evidence.

world-runtime is for teams working in environments where state, events, replay, simulation, and policy all need to stay explicit because the cost of getting decisions wrong is real.

This page brings together the current public portfolio: eleven implemented proof paths and one host-bound overlay track. It gives you the shape of what's here while GitHub remains the source of truth.

Why this category matters

This is for operating real worlds, not just routing work.

Most workflow software is built to push work from step to step. world-runtime is built for systems where entities, relationships, events, projections, proposals, branches, and governed decisions need to remain explicit, testable, and replayable.

Typical workflow stack
world-runtime
Moves work between people and screens
Models the operating world, then governs action inside it
State is often implicit across forms, tickets, and service boundaries
World state is explicit and rebuildable from append-only events
History shows activity, but not always causality or consequence
Replay and projection explain how current state was reached
Approvals are often process steps
Policy can deny, warn, or require approval before mutation

Current public portfolio

See the public portfolio at a glance

DomainDescription
Supply network planning and disruption response
Model suppliers, routes, inventory, and delays as explicit world state so teams can replay history, compare reroute strategies, and keep what-if analysis separate from canonical state.
Safety-constrained air traffic coordination
Use policy-first controls where unsafe decisions must be denied or escalated and where evidence matters before action can be approved.
Governed semantic coherence and meaning change
Model terms, definitions, and mappings as explicit world state so meaning changes stay reviewable, provenance-visible, and policy-governed.
Power-grid balancing and contingency response
Model balancing authorities, transmission constraints, and staged interruption decisions as explicit world state so cascading infrastructure choices remain simulation-backed and reviewable.
Cross-agency city operations incident coordination
Model agencies, transit, utilities, and response zones as one shared operating world so civic incidents stay coordinated, approval-visible, and evidence-backed.
Regulated lab science evidence and release governance
Model sample batches, instrument runs, evidence chains, and deviation review as explicit world state so regulated releases stay provenance-visible and approval-governed.
Market micro exposure control and inventory pressure
Model venues, books, inventory pressure, and risk limits as explicit world state so market interventions remain reviewable, simulation-backed, and desk-governed.
Multiplayer shared-state reconciliation and fairness control
Model contested match state, shard authority, and rollback bounds as explicit world state so simultaneous player intents stay reviewable, simulation-backed, and live-ops governed.
Autonomous vehicle motion planning and supervised intervention
Model vehicles, road segments, occluded hazards, and teleassist decisions as explicit world state so dynamic motion plans stay safety-limited, reviewable, and simulation-backed.
Multi-agent AI coordination, branch review, and governed delegation
Model agent teams, shared context, and review boards as explicit world state so delegated plans stay reviewable, simulation-backed, and policy-governed under branching pressure.
Open-agent-world shared commons intervention and stewardship
Model world zones, agent cohorts, governance beacons, and resource pools as explicit world state so emergent conflicts stay intervention-ready, approval-visible, and bounded by explicit oversight.
digital-twin-miniOverlay
A host-bound overlay track that keeps the same runtime artifact shape while attaching digital twins to other domains instead of standing alone as a showcase scenario.

Public proof paths

Each path proves the runtime under a different kind of pressure.

The cards below summarize the strongest public tracks in the repo today, with direct paths into the relevant playbook and adapter materials.

Public proof path

Supply network planning and disruption response

Public

Model suppliers, routes, inventory, and delays as explicit world state so teams can replay history, compare reroute strategies, and keep what-if analysis separate from canonical state.

What it proves
Replay, projection, and simulation tradeoffs
What it has to handle
Medium policy, high branching, and consequence tracing
Current state
Implemented proof path

Public proof path

Safety-constrained air traffic coordination

Public

Use policy-first controls where unsafe decisions must be denied or escalated and where evidence matters before action can be approved.

What it proves
Hard safety outcomes under time pressure
What it has to handle
Very high policy with explicit deny and require-approval paths
Current state
Implemented proof path

Public proof path

Governed semantic coherence and meaning change

Public

Model terms, definitions, and mappings as explicit world state so meaning changes stay reviewable, provenance-visible, and policy-governed.

What it proves
Governed meaning changes with provenance and semantic alternatives
What it has to handle
High governance with explicit deny and require-approval paths
Current state
Implemented proof path

Public proof path

Power-grid balancing and contingency response

Public

Model balancing authorities, transmission constraints, and staged interruption decisions as explicit world state so cascading infrastructure choices remain simulation-backed and reviewable.

What it proves
Cascading simulation with least-bad contingency tradeoffs
What it has to handle
High infrastructure governance with explicit deny and require-approval paths
Current state
Implemented proof path

Public proof path

Cross-agency city operations incident coordination

Public

Model agencies, transit, utilities, and response zones as one shared operating world so civic incidents stay coordinated, approval-visible, and evidence-backed.

What it proves
Cross-agency coordination with resident-impact tradeoffs
What it has to handle
High coordination with explicit deny and require-approval paths
Current state
Implemented proof path

Public proof path

Regulated lab science evidence and release governance

Public

Model sample batches, instrument runs, evidence chains, and deviation review as explicit world state so regulated releases stay provenance-visible and approval-governed.

What it proves
Evidence integrity with supervised release alternatives
What it has to handle
High provenance with explicit deny, warn, and require-approval paths
Current state
Implemented proof path

Public proof path

Market micro exposure control and inventory pressure

Public

Model venues, books, inventory pressure, and risk limits as explicit world state so market interventions remain reviewable, simulation-backed, and desk-governed.

What it proves
High-intensity exposure controls with deny, allow, and review alternatives
What it has to handle
High event intensity with explicit deny, warn, and require-approval paths
Current state
Implemented proof path

Public proof path

Multiplayer shared-state reconciliation and fairness control

Public

Model contested match state, shard authority, and rollback bounds as explicit world state so simultaneous player intents stay reviewable, simulation-backed, and live-ops governed.

What it proves
Shared-state reconciliation under concurrent update pressure
What it has to handle
High contention with explicit deny, warn, and require-approval paths
Current state
Implemented proof path

Public proof path

Autonomous vehicle motion planning and supervised intervention

Public

Model vehicles, road segments, occluded hazards, and teleassist decisions as explicit world state so dynamic motion plans stay safety-limited, reviewable, and simulation-backed.

What it proves
Motion-safety interventions under occlusion and teleassist pressure
What it has to handle
Hard physical safety with explicit deny, warn, and require-approval paths
Current state
Implemented proof path

Public proof path

Multi-agent AI coordination, branch review, and governed delegation

Public

Model agent teams, shared context, and review boards as explicit world state so delegated plans stay reviewable, simulation-backed, and policy-governed under branching pressure.

What it proves
Delegated coordination under branch-review pressure
What it has to handle
High governance with explicit deny, warn, and require-approval paths
Current state
Implemented proof path

Public proof path

Open-agent-world shared commons intervention and stewardship

Public

Model world zones, agent cohorts, governance beacons, and resource pools as explicit world state so emergent conflicts stay intervention-ready, approval-visible, and bounded by explicit oversight.

What it proves
Shared-world governance under emergent conflict pressure
What it has to handle
High intervention pressure with explicit deny, warn, and require-approval paths
Current state
Implemented proof path

Overlay track

digital-twin-mini

A host-bound overlay track that keeps the same runtime artifact shape while attaching digital twins to other domains instead of standing alone as a showcase scenario.

Host-bound

What it lets you build

Host-bound overlay twins attached to other domains

What it proves

Overlay simulation and host binding

What it has to handle

Overlay coordination across host domains with runtime-authoritative binding

Current state

Implemented; host-bound across power_grid and city_ops

How the public tracks are packaged

A shared bundle contract keeps every public track legible, comparable, and runnable.

The standard non-overlay bundle keeps public scenario tracks comparable across domains while leaving room for supplemental proof artifacts when a domain needs deeper evidence.

Standard track artifacts

README.mdentities.jsonrelationships.jsonevents.jsonproposal.jsondecision.jsonsimulation.jsonpolicy.jsonrule.jsonprojection.json

What the overlay adds

host_bindings.json

The overlay exception keeps the same runtime artifact set and adds host bindings so a twin can attach to another domain without becoming its own standalone proof path.

Support posture

Some surfaces are built for supported use. Others are still best-effort by design.

Stable and supported

  • App Server protocol
  • Public API /v1
  • Python SDK starter
  • Persistence migrations
  • Extension contracts

Experimental or best-effort

  • Starter templates and exploratory examples unless explicitly promoted
  • Future adapter additions or post-promotion follow-up work unless explicitly promoted
Read the support policy

How to evaluate and get moving

Start with the path that fits how you work.

The site is intentionally curated. The repository is the source of truth. These paths point each audience to the right starting set of materials.

Where to start

Start with the framing. Go deeper when the fit is real.

If you are still evaluating fit, begin with the docs hub and the adapter portfolio. If the category already clicks, go straight to the developer quickstart and the supported API surfaces.