Showcase

This section is the public, fixture-driven view of nous. Every page in the showcase carries a fidelity badge so a reader can tell at a glance whether a number came out of a calibrated filter, a real subsystem with no filter on top, a stub returning a constant, or a planned surface that has not landed yet. The legend lives in Fidelity.

The four pages below are the heart of the showcase. The scenario gallery is regenerated on every docs build by scripts/gen_showcase_telemetry.py; the rest are hand authored and reviewed when the code they describe changes.

Page	What it shows	How it stays current
Fidelity	The badge enum and per-subsystem mapping.	Reviewed when a subsystem's substance level changes.
State machine	Mission posture FSM with every transition.	Reviewed when src/nous/state/machine.py changes.
Capability matrix	Per-subsystem implementation, estimator, and model card.	Reviewed alongside the backlog.
Scenario gallery	Per-scenario telemetry, sparklines, timelines.	Regenerated in CI from scenario YAML.

The showcase is deliberately not a marketing page. The simulator is pre-1.0; the development line carries the L1 subsystem rollout (ten subsystems, nine estimators) but the self-model layer remains a placeholder, the scenario injectors are not yet wired, and no number has yet been compared against measured device traces. See the 2026-05-23 audit, the 2026-05-21 review, and AUDIT.md for the substance findings. The fidelity badges exist to keep the showcase honest about that. ADR 0017 records the decision behind this section.

If you are looking for the live reference instance, it remains private and gated; see docs/deployment.md for the deployment pattern. The live VM tracks origin/main via auto-update; if main lags the development line, the live MCP serves an older surface than this showcase claims (see audit-2026-05-23 §4 for the most recent probe).