---
title: Spending caps
description: Configure daily, monthly, or lifetime spending caps and overrun policies in the Zorveus dashboard.
---

Spending caps define the base virtual allowance for your organization, app connections, and product users. A cap acts as a spend ceiling, not a prepaid deposit.

You configure and monitor spending caps directly in the [Zorveus Dashboard](https://app.zorveus.com).

---

## Configuring caps in the dashboard

To set up or update spending caps for an application:

1. Open the [Zorveus Dashboard](https://app.zorveus.com) and navigate to **Apps**.
2. Select your application and open **Connection settings**.
3. Under **Spending caps**, configure:
   - **Cap amount**: Maximum reference-priced virtual spend permitted during the period (such as `$50.00`).
   - **Period**: Choose `daily`, `monthly`, or `lifetime`.
   - **Allowance overrun policy**: Choose how Zorveus handles requests that exceed remaining allowance.
4. Save your changes. New limits take effect immediately across all edge gateways.

The dashboard displays a live progress bar for each active cap showing current period spend, remaining base allowance, and the next reset timestamp.

---

## Cap scopes

Zorveus supports three cap levels:

- **Organization cap**: Configured in **Settings** -> **Billing**. Restricts aggregate AI spend across all apps and keys in the organization.
- **App-connection cap**: Configured in **Apps** -> **Connection settings**. Restricts total spend through that connection and its API keys.
- **Product-user cap**: Sets an individual customer base allowance on that key. You can monitor individual customer spend under **Product Users** in the dashboard.

When multiple caps apply, every limit must have sufficient remaining allowance for a request to proceed. Lowering a cap does not reverse or alter settled usage.

---

## Allowance formula and consumption order

When an inference key has a configured cap, total spendable allowance combines the remaining base cap with any active product-user credits:

```text
remaining allowance = remaining base allowance + available product-user credits
```

The remaining base allowance is:

```text
remaining base allowance = max(0, cap amount - completed period virtual spend - active reservations)
```

### Base cap is consumed first

When requests arrive, Zorveus consumes allowance in this order:

1. **Remaining base allowance** on the inference key.
2. **Available product-user credits** (oldest expiring grants first).
3. **Uncovered virtual spend** (in non-blocking track-only mode or request overruns).

```text
Configured monthly cap:        $10.00
Virtual spend so far:           $8.00
Remaining base allowance:       $2.00
Available user credits:         $5.00

Incoming request estimate:      $3.00
Base allowance reserved:        $2.00
Credits reserved:               $1.00
Remaining credits after call:   $4.00
```

The base cap covers the first $2.00. The excess $1.00 is deducted from user credits.

---

## Uncapped connections

If no cap is configured on an app connection in the dashboard:

- Virtual allowance is unlimited.
- Inference requests pass through without allowance rejections.
- Virtual spend is recorded in your dashboard analytics.
- Zero product-user credits are consumed.

Product-user credits extend a configured cap. They do not restrict spending when an app connection is uncapped.

---

## Allowance overrun policies

In your app connection settings in the dashboard, the **Allowance overrun policy** controls what happens when a model response exceeds remaining allowance:

| Policy | Dashboard setting | Behavior on active request | Excess virtual spend |
| :--- | :--- | :--- | :--- |
| `limit_output` (default) | Limit output tokens | Lowers provider max tokens to the affordable amount | Records unexpected variance as uncovered |
| `allow_overrun` | Allow overrun | Sends full client or safe output limit | Records excess as uncovered virtual spend, forgiven on next reset |
| `carry_overrun_forward` | Carry overrun forward | Sends full client or safe output limit | Records excess as debt, repaid from next credit grant or cap reset |

If an enforced request exceeds remaining allowance before calling the provider, the gateway rejects the call with HTTP `403 Forbidden` (`zorveus_product_user_allowance_insufficient` or `zorveus_cap_exceeded`). The dashboard logs the rejection in your usage events.