A workbench built for governance, not just visibility.

Every AI component in ModelAIr carries an assertion level. This is not a label — it is a governance constraint that defines what the component is actually permitted to do in the system.

Signal: the component produces outputs that are observed but not acted on autonomously. Advisory: the component produces outputs that inform human decisions. Authoritative: the component produces outputs that trigger downstream action without human approval.

The distinction matters legally, operationally, and architecturally. ModelAIr makes it explicit and exportable. The assertion level is part of the governance record from the moment the component is placed.

Authority boundaries define the perimeter of autonomous AI action within your system. ModelAIr lets architects model this explicitly — drawing a boundary is a first-class design action, not an afterthought.

Within the boundary: AI components may act autonomously. Outside it: a human checkpoint is required before action proceeds. The boundary is visible in the diagram, encoded in the governance record, and exportable to compliance artefacts.

This is the question regulators and risk teams are asking. ModelAIr gives architects the mechanism to answer it with precision rather than inference.

In most architecture tools, connections are lines. In ModelAIr, every connection between components is a typed edge carrying structured properties: transport protocol, authentication method, data sensitivity classification, and reliability characteristics.

A connection between a model and a data store is not a line. It carries the protocol (REST, gRPC, message queue), the authentication method (API key, OAuth, mTLS), the classification of data in transit, and the reliability model (sync, async, best-effort, guaranteed).

Edge semantics are part of the governance output. The connection is as much a governed object as the components it connects.

There is no export step. There is no documentation layer. The design workspace in ModelAIr is the governance record. As architects design, the governance model is built automatically — structured, queryable, and ready for audit.

The output can be queried directly: which components carry authoritative assertion levels? Which data stores handle sensitive data? Which human checkpoints are in the critical path? These are answerable from the model — not from a PDF.

When regulators ask for AI system documentation, ModelAIr gives organisations something they can actually stand behind. Not a diagram. A governed, structured record of how the system is built and what it is permitted to do.