Watcher Profiles vs Deployments vs Slots
Three related but distinct concepts your plan's capacity is measured in.
Watcher Profile — a reusable AI moderator identity/persona (a `virtual_moderators` row). Creating one does not by itself consume a Watcher slot.
Deployment — assigning a Watcher profile to a specific Community. A Watcher isn't live in a Community until it's deployed there.
Watcher Slot — one ACTIVE (enabled) Watcher-to-Community deployment. Your plan's total slot allowance is checked against the count of active deployments, not the count of distinct Watcher profiles.
Example: the same Watcher deployed to two Communities uses two slots. Two different Watchers deployed to the same Community also use two slots. Pausing a deployment frees its slot immediately; reactivating re-checks your current capacity.
Regardless of your plan's total slot allowance, any single Community can hold at most 5 active Watcher deployments.
A Watcher deployed into a private Escalation Group is the one exception — that deployment never consumes a Watcher slot, since the Escalation Group is internal ops infrastructure, not a customer-facing Community. See "Watcher Slots and Billing" for the full formula and worked examples.