The vocabulary of Aerius cognitive runtime architecture.

A system-level structure that coordinates memory, observation, reasoning, identity, knowledge, action, expression, traceability, and governance.

The AI model is one component inside a broader architecture, not the entire system.

A runtime layer that can expose what it did, why it did it, what evidence it used, and which policy applied.

An operational layer that governs a cognitive function such as observation, memory, reasoning, domain logic, action, or expression.

Connecting a structured component, protocol, domain capsule, tool, verifier, or reasoning object into the active cognitive runtime.

A structured container for observation results, source metadata, evidence level, policy, risk signals, and trace information.

Rules for what should be observed, from which source, at what mode, duration, evidence level, and constraint set.

Observation guided by the evidence needed for reasoning, verification, risk handling, or decision support.

A mechanism for assigning cognitive priority, reasoning depth, urgency, risk handling, confidence handling, and resource allocation.

Structured reasoning stacks that separate, coordinate, compare, or synchronize reasoning processes.

Reasoning about which reasoning strategy, process, or policy should be used.

Memory associated with a specific person, entity, role, organization, or identity anchor.

Reasoning that accounts for relationship-specific context such as trust, familiarity, role, safety, boundaries, and expectation gaps.

Trust evaluated by domain or context rather than as one global score.

Extracting, structuring, validating, promoting, demoting, quarantining, and governing knowledge derived from humans or interactions.

An operational unit of domain-specific knowledge, reasoning protocols, rules, tools, verifiers, constraints, risk profiles, and authority boundaries.

Domain knowledge structured to guide reasoning, verification, tool use, constraints, and policy inside a runtime.

A previous cognitive processing outcome can be reused only when current task, context, evidence, time, risk, and policy conditions justify reuse.

A structured representation of what a task is, what context it depends on, what evidence it uses, and what constraints apply.

The system's structured representation of cognitive and emotional posture for communication and behavior.

A structured representation of how the system should express an internal cognitive state or communication goal.

A structured representation of what the system intends to communicate, do, ask, block, review, or execute.

An intended action described at a meaningful level before hardware-specific, API-specific, or execution-specific implementation.

An output, response, expression, or action that has passed through the relevant runtime context, policy, risk, evidence, and review constraints before being delivered.