StewardStatusResponse

Whether the steward exists and what serves it right now.

enabledEnabled (boolean)required

Whether intelligent-fabric mode is enabled in Settings.

presentPresent (boolean)required

Whether a steward placement currently exists in state.

readyReady (boolean)

Whether every runner of the steward placement reports Ready: present-but-not-ready means the model is still downloading or loading, and chat requests will queue or fail until ready.

Default value: false

steward_model object

Model id of the steward brain, when present.

anyOf

• string
• null

string

instance_id object

The steward instance id, when present.

anyOf

• string
• null

string

desired_model object

Preferred brain currently being prepared, or the serving brain when no transition is active.

anyOf

• string
• null

string

transitionTransition (string)

Controlled best-brain lifecycle: idle, prestaging, replacing, or repairing after a placement disappeared.

Possible values: [idle, prestaging, replacing, repairing]

Default value: idle

progress object

Aggregate prestaging completion ratio when measurable.

anyOf

• number
• null

number

Possible values: >= 0 and <= 1

stateState (string)

One-word lifecycle summary derived from the other fields plus canary history: 'disabled' (mode off), 'downloading' (placed, weights still staging), 'starting' (placing or loading), 'ready' (serving, clean canary), 'degraded' (serving, but the liveness canary has an outstanding failed probe). The boolean fields remain authoritative; this is the renderable summary.

Possible values: [disabled, downloading, starting, ready, degraded]

Default value: disabled

StewardStatusResponse

{
  "enabled": true,
  "present": true,
  "ready": false,
  "steward_model": "string",
  "instance_id": "string",
  "desired_model": "string",
  "transition": "idle",
  "progress": 0,
  "state": "disabled"
}