Routing calibration
Calibration lets you bias which agents the planner favors, without ever making an unavailable or incapable agent eligible for a task.
AgentDock Pro
Routing calibration is part of AIgency, available only in AgentDock Pro.
Task routing scores agents against a task using a static capability matrix. Calibration sits on top of that scoring as a bias dial — it changes which eligible agent tends to win, not which agents are eligible in the first place.
Presets
- Balanced — the default; preserves the base routing behavior with no bias applied.
- Claude Heavy, Codex Heavy, Visual/UI Heavy, Efficiency — each nudges scoring toward agents suited to that emphasis.
- Custom — set per-category weighting yourself rather than using a preset.
What calibration can't do
Note
Calibration reorders which usable agent the router prefers. It never makes an otherwise-ineligible agent eligible — an agent that fails the task's capability or availability constraints stays excluded regardless of how calibration is set.
Calibration is stored as a Settings-level default (aigencyRoutingCalibration — see AIgency settings) and applies to new planning decisions; it doesn't retroactively reassign already-planned tasks.