STACK
MENU
DOCS / API REFERENCE / USAGE

Usage API

GET /v1/usage is the Console's canonical usage, spend, and earnings response. It is operator-scoped and uses UTC windows.

bash
curl "https://api.getstack.run/v1/usage?since=2026-08-01T00:00:00Z&until=2026-09-01T00:00:00Z" \
  -H "Authorization: Bearer $STACK_TOKEN"

Consumption

  • consumption.activity.total counts every audit event only when total_coverage is complete. Partial coverage makes the total a lower bound.
  • consumption.activity.total_coverage states whether the activity total is complete or partial.
  • consumption.actions counts only actions charged against the monthly action allowance.
  • allowance_coverage states whether the historical month has a durable tier snapshot.
  • detail_coverage states whether by_kind is complete, partial, or unavailable.
  • Legacy passports, proxy_calls, dropoffs, and skill_invocations fields have limit: null. They are breakdowns, not separate quotas.

Seats, allowances, and inventory

Seats include the owner. A billing-period response reports current-month pass-through allowances for L3 classifications, SMS, and email. A custom time window returns passthrough: null because the historical counters are not durable. Agents, service connections, published skills, and inbound webhook endpoints are current point-in-time counts under unlimited. For a custom window, agent and published-skill rows also include created_in_window. Fair-use egress and agent-registration ceilings are published separately.

Money

spend_cents reports actual STACK wallet spend by bucket. Top-ups, refunds, and adjustments are wallet movements, not product spend. LLM provider costs remain separate from STACK fees and wallet debits. The response also includes wallet balance and unpaid and lifetime publisher earnings.

An old month with unknown allowance coverage never borrows the current plan limit. Render the coverage field instead of inventing a historical quota.

stack | Docs