NeuroFabric · Pedagogical model
NeuroGrid
An interactive mental model for how the architecture program talks about policy-bounded movement, governed action, and ledger-style traces. It is intentionally minimal and abstract — not a simulator of a full runtime, not evidence of production scale, and not a substitute for reading the architecture reference.
Scope. This page does not demonstrate customer-ready deployment, enterprise certification, or feature parity with any real control plane. It only helps intuition. For current deployment conversations, Atlas is the appropriate path (see below).
Interaction mode
Scenario preset (legend labels only — same mechanics)
Concept map (demo ↔ architecture language)
- • Policy — Red regions: moves into these cells are rejected, analogous to a policy gate or disallowed capability. The grid edge plays a similar role.
- • Execution / governed action — Each accepted move is a discrete step on a path. Illegal moves do not occur; only allowed transitions extend the trace (simplified from a real planner → orchestrator pipeline).
- • Ledger, review, replay — The trail behind the marker stands in for an ordered execution record. Replay (where available) re-walks that order for intuition, not a guarantee of byte-identical production replay.
What the grid is illustrating
- • Trail — Treat as a stand-in for an execution ledger: append-only along your play, visible for inspection.
- • Goals — Green cells resemble successful completion of an allow-listed “tool” or milestone inside the policy — still abstract, not API-accurate.
- • Replay — Press P to replay your recorded order of moves.
In-market product
For current deployment conversations, see Atlas .