# How usage billing works

> How web services, instances, game servers and databases bill by the second, what a month includes, and what the Paid plan and first deposit are for.

Source: https://www.coritan.com/docs/billing/usage-billing/

In the dashboard:

- /dashboard/billing: https://www.coritan.com/dashboard/billing
- /dashboard/billing/usage: https://www.coritan.com/dashboard/billing/usage

Coritan bills the web services and sites on App Deployment, Cloud Compute instances, and game servers and databases by what each uses. It measures use by the second and takes the cost from your credit each hour, or puts it on your monthly invoice. A service has no price of its own, so a small one that stays inside the amounts every month includes can cost nothing.

This page covers what is billed by use, what a month includes, and what you need before you order. [Usage rates](/docs/billing/usage-rates/) lists every rate and included amount, and each product has its own page with a worked example.

## What is billed by use {#what}

| You order | You choose | What is billed | You need |
| --- | --- | --- | --- |
| A web service or a hybrid site on App Deployment | A source, a name and a location | The memory and CPU its instances use, the requests they answer, the data they send, builds and logs | The first deposit, on a prepaid account |
| A static site on App Deployment | A source, a name | Builds and logs only | Nothing |
| A Cloud Compute instance | A size | Memory while it runs, disk while it exists, CPU it uses and the data it sends | The Paid plan, and the first deposit on a prepaid account |
| A game server or a database | A size | The same as an instance | The Paid plan, and the first deposit on a prepaid account |

[How App Deployment is billed](/docs/managed-containers/how-billing-works/), [How Cloud Compute is billed](/docs/cloud-compute/how-billing-works/) and [How game servers and databases are billed](/docs/managed-containers/how-servers-are-billed/) explain each product.

A web deployment has no plan and no size. An instance, a game server and a database keep their sizes (RX for instances, CX for game servers and databases), which set the vCPUs, memory and disk, but a size does not set a price.

A service you ordered before usage billing keeps the price it has until you move it. [Switch a service to usage billing](/docs/billing/switch-to-usage-billing/) moves one of yours. A service an organization sells to its customers keeps the price the organization set.

## What the meters measure {#meters}

Memory
: A web deployment pays for the memory its running instances hold. A larger memory limit costs nothing until an instance uses it. A server pays for its memory size for the seconds it runs.

CPU
: The processor time a service uses, never the processor time it could use. An hour with no CPU reading from the host is not billed for CPU.

Disk
: A server's disk, for as long as the server exists. A stopped server keeps paying for its disk and nothing else. A web deployment has no disk meter.

Requests and data out
: A web deployment pays for the requests an instance answers. Every service pays for the data it sends to visitors, by the region of its location. A static file, a cache hit and a page the edge writes itself cost nothing.

Builds and logs
: A web deployment pays for the minutes its builds run and for the log data its instances write.

Each [rate](/docs/billing/usage-rates/#rates) is per unit of a meter, such as a GiB-hour or a million requests.

## What each month includes {#included}

Every month includes some use at no charge, and it starts again on the 1st (UTC). The amounts come in two pools, and each pool is shared by all your services of that kind.

Web pool
: Every account has it, on every plan. It covers your web services, hybrid sites and static sites.

Server pool
: Your instances, game servers and databases draw on it while your account has the [Paid plan](#paid-plan). Without the Paid plan, its amounts are zero and every unit is billed.

Your use draws an amount down hour by hour. Past it, each unit is billed at its rate. Nothing is held back or stopped at an amount. The amounts are in [Usage rates](/docs/billing/usage-rates/#included).

## The Paid plan {#paid-plan}

The Paid plan is what an account needs to order a Cloud Compute instance, a game server or a database. It costs the same for the whole account, however many servers it runs. A web deployment needs no plan.

- Pro, Business and Enterprise include it. Their accounts have nothing to add.
- A Free account adds it. It is an account add-on named Paid plan, which renews monthly, quarterly or yearly ([Usage rates](/docs/billing/usage-rates/#paid-plan)). Order it on its own with the API, as [Add capacity to a service](/docs/billing/add-ons/#add-the-paid-plan) describes.
- An order without it is refused. The API answers `402` with the code `paid_plan_required`, and names the add-on and the plan that includes it.

## What you need before you order {#deposit}

A prepaid account must have added $10.00 of credit in total before it runs an instance, a game server, a database or an instance of a web deployment, by default. The deposit stays in your account as credit and pays your first hours. An account billed by invoice needs no deposit, and neither does a static site. [How hourly billing works](/docs/billing/hourly-billing/#unlocking-hourly-billing) explains the rule.

Until the deposit is in, the order answers `403`, and the quote says a deposit is required.

## How you pay {#paying}

A prepaid account pays from its credit every hour. The credit's history shows each charge as `Compute usage for service #311`, with the service's number. When your credit runs short, Coritan tries your [automatic top-up](/docs/billing/automatic-top-up/) first. If that fails or is off, previews sleep and instances added under load stop, and the service is suspended with the reason `Insufficient credit balance`.

An account billed by invoice gets one invoice a month, once the month has been counted in full. It has a line for each meter of each service, such as `Memory above the included amount`, with the quantity and the rate. A meter that comes to less than half a cent waits for the next invoice. A service you cancel gets a final invoice for what it still owes.

## Stopped, sleeping and free {#free}

- A stopped server costs its disk only. Its memory and CPU are not billed while it is off.
- A sleeping web instance costs nothing. Turn on **Sleeps when idle** when you order a deployment, or [set it later](/docs/managed-containers/locations-and-scaling/).
- A static site costs nothing to run. It has no instance, so only its builds and logs can be billed, and they draw on the web pool.
- A free server is never billed. It stays free on the brands that sell it, and so does every boost it gets ([Free servers](/docs/managed-containers/free-servers/)).

## Cap what a month can cost {#limits}

A spend limit stops a month from costing more than you chose. An account limit counts every service of the account, instances included, and a deployment can have a limit of its own. At 100%, **Hold back the extras** sleeps previews and stops the extra instances a deployment started under load, and an account limit also refuses new orders of services billed by use. **Stop every instance** stops everything that runs, and suspends your Cloud Compute instances too. We email you at 50%, 80% and 100% of a limit. [Spend limits](/docs/managed-containers/how-billing-works/#spend-limits) has the steps.

## Read what you used {#read}

- In **Billing**, the **Usage** tab shows every service on the account, split by product, with the free amounts left and where the month is heading ([See your compute usage and forecast](/docs/billing/see-your-usage/)).
- On a deployment, the **Billing** tab reads **Billed by use** and shows its rates, what it used and cost this month for each meter, the free amounts and its spend limits. The **Usage** tab shows what it used and cost this month and in earlier months, and how much of each free amount your account has used.
- The API lists a month's use for every service, with a line for each meter, in [`GET /api/v1/billing/usage`](/docs/api/reference/client/billing/#op-get-api-v1-billing-usage). An instance, a game server or a database has no deployment, so its entry has a `null` `uuid` and its `service_id`.

```bash
curl "https://api.coritan.com/api/v1/billing/usage?month=2026-10" \
  -H "Authorization: Bearer $CORITAN_TOKEN"
```

Each line has the `meter`, its `unit`, the `quantity` used, what the month's amounts `included`, what was `billable` and the `amount` it cost.

## Related

- [Usage rates and included amounts](/docs/billing/usage-rates/)
- [See your compute usage and forecast](/docs/billing/see-your-usage/)
- [Switch a service to usage billing](/docs/billing/switch-to-usage-billing/)
- [How hourly billing works](/docs/billing/hourly-billing/)
- [Add capacity to a service](/docs/billing/add-ons/)

## API

- `GET /api/v1/billing/usage`: Deployment usage (https://www.coritan.com/docs/api/reference/client/billing/#op-get-api-v1-billing-usage)