LOADING
Core explanatory content is server-rendered. This state is only for a route segment that is still resolving.
TRACES
Trace an AI response back to its Behavior Point, revision set, authority layers, signals, projection, model target, and immutable execution receipt.
DEFINITION
A Vira runtime trace is explanation evidence for one execution. It connects the served result to exact behavior authority instead of showing only a prompt snapshot or generic application log.
HOW IT WORKS
Resolve the explicit Behavior Point for the current project, application, workspace, user, and session context.
Record the exact active revision set and deterministic composition identity.
Select the bounded behavior capsule using the Behavior Point and trusted signals.
Record which authority layers and modules applied and which were skipped.
Record canonical or governed optimized projection identity together with the model target.
Persist the runtime receipt so later review can explain the execution without guessing from a moving latest state.
PRODUCT PROOF
These fields describe the product evidence Vira preserves for this part of the lifecycle.
CONTROL BOUNDARIES
The trace points to exact revisions, composition, projection, and model target rather than a mutable 'latest' behavior.
Explanation includes which layers were eligible, rejected, delegated, conflicting, or otherwise skipped.
Signals retain source authority so a model-assisted extraction cannot impersonate a trusted application or developer fact.
Trace content depth can follow configured privacy modes while operational identity and governance evidence remain available.
LIMITATIONS & FAILURE STATES
DEVELOPER HOOK
The SDK flow preserves the resolve-issued identity and lets the application attach the model result without calculating a behavior revision client-side.