module/ai-runtime-ui

AI Runtime UI

Surfaces for the on-device intelligence layer.

Whispers, intent tiles, confidence indicators, and inline reasoning shells. Designed for local-first inference where latency is sub-100ms and tokens stream into the UI.

State Logic

Reactive surfaces this module exposes to the runtime.

statusenum

idle | thinking | streaming | done | refused.

confidence0..1

Drives indicator ring and color.

tokensPerSecnumber

Live throughput meter on streaming surfaces.

modelIdstring

Local model identifier. Always shown, never hidden.

Visual Variants

Allowed rendering modes. No ad-hoc styles outside this set.

whisper.inline

One-line assistance under inputs.

whisper.capsule

Floating suggestion, dismissible.

intent.tile

Navigation prediction in sidebar.

reason.shell

Expandable chain-of-thought block.

refusal.card

Scoped denial with policy reason.

Motion Tokens

Deterministic timing curves. No custom easing per component.

tokenvalueuse
ai.stream60ms ease-out per tokenToken-by-token reveal.
ai.think1200ms ease-in-out loopPulse while inferring.
ai.settle240ms cubic-bezier(.2,.8,.2,1)Final layout on done.

Isolation Rules

Hard boundaries enforced at the module edge.

01

Model identity is non-removable from any AI surface.

02

Confidence below 0.4 must render the indicator in warn tone.

03

Whispers never block input. Intent tiles never auto-execute.

04

No AI surface may write to wallet or agent state directly.

// end module /ai-runtime-ui