DESKTOP 3D · WEBXR · ACCESSIBLE FALLBACKS

SEE THE ASSURANCE STRUCTURE IN SPACE.

10 working simulations turn the supplied research into inspectable software. Every scene runs in a browser, renders in native WebGL, supports immersive WebXR where the device allows it, and exposes equivalent table and text views.

Open the first simulation View proof of function Research-to-code traceability →

10 IMPLEMENTED SCENES

FROM REPORT REQUIREMENT TO RUNNING SOFTWARE.

Each simulation has deterministic server logic, a browser 3D scene, WebXR support, accessible alternatives, API endpoints, proof invariants, and direct links to the research that shaped it.

3DWEBXRAPITEXT

Accountability Evidence Room

Walk through the complete evidence-to-outcome structure hidden behind a simple approval record.

Purpose: Demonstrate that a recorded human click does not prove independent judgment or erase upstream contributing factors.
Research sources: 3
Proof invariants: 3

Run Evidence Room

3DWEBXRAPITEXT

Decision Provenance Replay

Replay an append-oriented synthetic history in three dimensions and create a clearly marked counterfactual branch.

Purpose: Demonstrate deterministic reconstruction without rewriting canonical history or equating replayability with correctness.
Research sources: 3
Proof invariants: 3

Run Provenance Replay

3DWEBXRAPITEXT

Dynamic Assurance and Change Impact

Inject a software, model, dataset, policy, interface, environment, compiler, evidence-age, or reviewer-authority change and observe claim-state propagation.

Purpose: Demonstrate conservative assurance states, defeaters, stale evidence, and renewed-verification duties.
Research sources: 3
Proof invariants: 3

Run Assurance Graph

3DWEBXRAPITEXT

Capability, Trust, Authority, and Permission Gates

Move an abstract request through technical capability, data, trust, authority, purpose, policy, and contextual-permission gates.

Purpose: Demonstrate that CAN does not imply MAY and that missing authority cannot be averaged away by favorable technical conditions.
Research sources: 3
Proof invariants: 3

Run Authority Gates

3DWEBXRAPITEXT

Uncertainty Field and Abstention Space

Separate model-output confidence from aleatoric uncertainty, epistemic uncertainty, evidence completeness, source independence, and distribution status.

Purpose: Demonstrate that a polished high-confidence output may coexist with high ignorance and should trigger abstention.
Research sources: 3
Proof invariants: 3

Run Uncertainty Field

3DWEBXRAPITEXT

Federated Data, Trust, Purpose, and Revocation

Move a signed synthetic data product across organizational boundaries while evaluating identity, integrity, semantics, disclosure, purpose, reuse, authority, and revocation separately.

Purpose: Demonstrate that access does not imply permission to reuse and that a trusted sender does not prove trustworthy data.
Research sources: 3
Proof invariants: 3

Run Federated Trust

3DWEBXRAPITEXT

Resilience Web: Degrade, Quarantine, Reconcile

Inject node failure, partition, latency, stale state, provenance invalidation, schema mismatch, authority expiry, or conflicting state into an abstract directed graph.

Purpose: Demonstrate connected, trusted, authorized, and permissible paths as nested but distinct sets.
Research sources: 4
Proof invariants: 4

Run Resilience Web

3DWEBXRAPITEXT

Modular Open Architecture and Vendor Swap

Swap a fictional component while testing interface ownership, conformance, data rights, software authorization, and rollback readiness.

Purpose: Demonstrate the difference between component performance and end-to-end thread interoperability without producing real procurement documents.
Research sources: 3
Proof invariants: 3

Run Open Architecture

3DWEBXRAPITEXT

Governance Lifecycle and Qualified-Human Gates

Move a fictional autonomous-system concept through technical feasibility, independent legal/policy review, institutional permission, bounded authority, and public accountability.

Purpose: Demonstrate lifecycle governance, explicit unknowns, independent review, and non-delegable human decisions without generating legal conclusions.
Research sources: 4
Proof invariants: 3

Run Governance Lifecycle

3DWEBXRAPITEXT

Autonomous Drone Control Assurance Lab

Control a synthetic three-drone inspection fleet inside a normalized 3D test volume while authority, geofence, battery, navigation integrity, communications, uncertainty, and human override remain explicit.

Purpose: Demonstrate bounded autonomous control, safe fallback, human override, decision provenance, and runtime-assurance gates without real hardware control, real-world coordinates, targeting, payloads, or weapon functions.
Research sources: 8
Proof invariants: 5

Run Drone Control

THE SOFTWARE CONTRACT

THE VIEWER AND THE ENGINE PROVE DIFFERENT THINGS.

The native WebGL/WebXR viewer proves that the spatial representation is interactive and hardware-capable. The deterministic Python engine proves declared state transitions, path-subset invariants, append-oriented history, authority gating, abstention, and safe degradation for the supplied synthetic inputs.

RUN

The browser sends a bounded JSON scenario to the same-origin Flask API.

DERIVE

Pure service functions calculate explicit conditions without a universal score.

VERIFY

Named invariants evaluate the result and return cryptographic digests.

RENDER

The 3D scene, table, text, event log, and proof record update together.