No slides, no mock data. Each button below POSTs/GETs a real a11oy API and renders the raw response. Replay re-derives a decision hash, tamper changes it, Λ collapses on a zeroed axis, the fleet reports honest peer health, and the mesh shows conserved cross-organ wires.
a11oy policy gates are deterministic pure functions. Re-execute a past decision from its inputs and get the same verdict + a canonical SHA-256 receipt hash — independently verifiable.
POST /api/a11oy/v1/receipts/replay
Replay the same action twice → identical hash (repeatable). Then tamper a single input (confidence 0.92 → 0.50) → the hash changes. Tampering is detectable, not hideable.
POST /api/a11oy/v1/receipts/replay ×3 (baseline · repeat · tampered)
Λ = ∏ xᵢ^wᵢ is zero-pinned: if any positive-weight trust axis hits 0, Λ collapses to 0 and the gate DENIES — regardless of the other axes. Compare a healthy score vs a single zeroed axis.
POST /api/a11oy/v1/lambda/score
The fleet probes every flagship organ's live health endpoint. Organs that answer report ok; ones that don't are shown honestly as unreachable — no fabricated green.
GET /v4/fleet
Five organs, one substrate. The mesh state lists the cross-organ wires (a11oy↔sentra, ↔rosie, ↔amaru, ↔vessels) and their live status. Honest labels distinguish in-process wires from cross-Space ones that are not yet brokered.
GET /api/a11oy/v1/mesh/state