Docs · Spend controls

Spend controls and billing behavior.

Model usage is the one genuinely variable cost of running agents. Qoren's design goal is that you are never billed for managed usage you did not pre-pay: usage draws from your plan's allowance, then from prepaid credits, and stops at zero.

The spend order

With the managed model key, every dollar of model usage follows the same path.

  1. 1Plan allowance firstPaid plans include a monthly managed-model allowance (the amount varies by plan). Usage draws it down first, and it resets with the billing period.
  2. 2Prepaid credits nextWhen the allowance is exhausted, or during a trial, usage draws from your prepaid credit balance.
  3. 3Hard stop at zeroWhen both are exhausted, managed model usage stops. Agents stay deployed and resume when you top up, switch to BYOK, or the allowance resets. There is no open-ended metered billing.

Prepaid credits

Credits are prepaid model-spend headroom, bought from Account & Billing with a $5-$100 slider. What you pay lands in your balance 1:1, pay $20, get $20 of credit.

  • Credits never expire and carry across billing periods.
  • They cover managed usage during a trial and beyond your allowance.
  • Checkout runs through Polar, the payment processor for all Qoren billing.

Bring your own key (BYOK)

Every plan, including the trial, can connect its own OpenRouter key instead of using the managed one. With BYOK you pay OpenRouter directly, any model is available, and the managed allowance and hard stop no longer apply, OpenRouter's own limits govern.

  • BYOK is OpenRouter-exclusive. The key is validated, stored encrypted, and applied across all your agents.
  • Disconnecting reverts to the managed key and re-enables the allowance and hard stop.
  • The card warns you when a connected key looks free-tier, so agents don't stall mid-run.

Watching spend

The Spending page meters usage where it happens, per agent and per task.

  • OpenRouter cost and tokens (in/out) per agent, with 24h / 7d / 30d / all-time windows.
  • Sessions broken down by source: scheduled runs, chat, Telegram, Slack, and the rest.
  • HTTP request counts per external API the agent calls, usage beyond the model bill stays visible.
  • An ≈ marker flags totals that include the provider's estimate before exact costs land.

For agencies

Because each client runs in separate environments, the Spending page's per-client grouping is also your per-client cost report, useful both for your own margins and for showing a client what their agents actually did.

Frequently asked questions

Managed model usage stops; nothing keeps billing. Agents stay deployed and resume when you top up credits, connect BYOK, or your allowance resets with the next billing period.

Keep reading

See it with a real agent.

The fastest way to understand the flow is to deploy a template and watch its first run.

Start free trial