Metering, cost and limits

Quotas and budgets

Cap tokens, spend or requests per person, team or service, with warnings before anyone hits the limit.

What it is

Cap input tokens, output tokens, spend or request count per person, team or service token, optionally per model. Windows are daily, weekly, monthly, or rolling 24 hours, 7 days or 30 days.

Why you want it

A runaway script or a curious intern should hit a limit long before the invoice does.

The Data science team has a $1,500 monthly budget. A warning goes out at 80 percent and again at 95 percent. At 100 percent new requests are refused with a message saying when the budget resets, while the request already in flight is allowed to finish. DATA SCIENCE · MONTHLY BUDGET $1,500 resets Nov 1 80% 95% At 80%: a warning to the team and its lead In the app, by webhook or by email. 95% sends another. At 100%: new requests are refused quota exceeded · resets 2026-11-01 00:00 UTC The person sees why, and when it ends. The request already running can finish Or cut it mid-stream, if you prefer a hard stop.

How it works

  • Warnings at configurable thresholds (default 80% and 95%) plus a breach alert
  • Breach behavior: let the in-flight request finish or cut it mid-stream
  • Refusals say when the limit resets
  • Team leads can manage their own team's quotas when an admin delegates it

In the product

Janus Quotas screen with demo data
Quotas

Per-person and per-team limits with live utilization, reset dates and what happens on a breach.

Actual Janus interface · Demo identities and synthetic usage

See it on your own network.

The Community edition is free for up to 25 people. The 30-day Business trial unlocks every Business feature.