Skip to main content

AI Usage & Credits

What Counts as AI Token Usage​

All AI calls routed through the platform's shared subscription are metered:

  • Chat conversations — every message sent and received in the chat panel
  • Task execution — AI calls made by automated tasks (refinement, workflow steps, code review)
  • Workflow AI — AI calls inside workflow instances

Both input tokens (your prompts, system context, attachments) and output tokens (the model's response) contribute to your token consumption figures shown on the Usage page.

How to Monitor Consumption​

Usage Page (Org Admins)​

Navigate to Usage in the sidebar to see your organization's AI token and cost consumption for the current window:

  • A progress bar shows tokens and estimated cost used vs. the configured reporting thresholds
  • If your burn rate is high enough that the reporting threshold is projected to be hit before the window ends, a projected date is shown (based on a 7-day trailing average)

These figures are informational — they reflect how much AI spend has occurred and power the warning banner. They do not enforce a hard block on their own.

Warning Banner​

When your organization's AI consumption reaches 80% of the configured reporting threshold, a non-blocking banner appears at the top of the Usage page. The banner shows:

  • Current consumption percentage
  • Projected exhaustion date (when applicable)

The banner can be dismissed. It reappears at the start of the next window.

API​

Org admins can query the AI usage reporting endpoint programmatically:

GET /api/orgs/:orgId/quota

Response:

{
  "scope": "global",
  "window": "monthly",
  "tier": "free",
  "limits": { "tokens": 1000000, "costUsd": 5.00 },
  "used": { "tokens": 342100, "costUsd": 1.71 },
  "remaining": { "tokens": 657900, "costUsd": 3.29 },
  "warnAt": 0.8,
  "isWarning": false,
  "windowResetsAt": "2027-01-01T00:00:00.000Z", // illustrative
  "projectedExhaustionAt": null,
  "allowed": true
}
Reporting only

The limits, remaining, and allowed fields in this response reflect the legacy AI-only reporting thresholds. They are not the enforcement decision — they do not indicate whether your organization's work will be blocked.

Super-Admin Controls​

Super-admins can view and override the AI token/cost reporting thresholds for all organizations at System → Quotas:

  • View each org's current AI token and cost consumption vs. configured thresholds
  • Change the reporting tier (free, paid, custom, unlimited)
  • Set custom token or cost thresholds
  • Add notes for auditability

All changes are logged in an audit trail with the admin's identity and a snapshot of what changed.

note

These controls adjust the reporting thresholds and warning banners shown on the Usage page. They do not control whether work is blocked.

Updating System-Wide Reporting Defaults​

Default reporting thresholds for new organizations are set in System → Settings:

  • global_subscription_default_monthly_token_limit (default: 1,000,000)
  • global_subscription_default_monthly_cost_usd_limit (default: 5.00)

Changing these defaults affects new organizations and organizations that do not yet have an explicit quota row. Existing quota rows are not retroactively modified.