> See the full index of documentation pages at https://docs.zettabyte.com/llms.txt.

# API keys

> Create and manage keys for the inference API.

Source: https://docs.zettabyte.com/docs/inference-api/api-keys

Last updated: 2026-10-02

API keys authenticate calls to your deployed inference endpoints (zCLOUD) and to zSTUDIO's model API. Manage keys on the [**API Keys**](https://app.zettabytecloud.com/API-key/list) page, which is in both workspaces on the console's switcher: **zCLOUD · AI Infrastructure Rental** and **zSTUDIO · MaaS via API**.

## Create a key

The key value is shown *once*, in a copy dialog at creation. Copy it then and store it securely; treat it like a password.

Select **Create API Keys** and set the fields below. If you don't see **Create API Keys**, you can't create a key from your view.

- **API Keys Name**: must be unique. A duplicate name returns "API key name already exists."
- **Permission Settings**: two boxes, **Inference Service: Reading** and **Inference Service: Write/Execute**. The boxes can't be selected, and a call to zSTUDIO's model API doesn't need either.
- **RPM** and **TPM**: RPM (requests per minute) and TPM (tokens per minute). The fields open filled with default values, and the form shows those defaults below the fields.
- **API Keys Validity**: **No Expiration** or **Customized Validity Period**.

You can create more than one API key, for example one per app or environment, and each shows its own usage and limits. A key isn't tied to a model: the model is set in each request.

## View your keys

Use the **API Keys** page to find a key and read its limits. Summary cards show total keys, active keys, calls in the last 30 days, and average delay.

The list has these columns: **Name**, **Secret Key**, **Status**, **Rate Limiting**, **Belonging Users**, **Creation Time** and **Recent Usage Time**. After creation, the **Secret Key** column shows a masked value, so the copy you made at creation is the one to use.

The **Rate Limiting** column shows each key's limits as RPM / TPM: requests per minute and tokens per minute.

For the errors a key can return, see [zSTUDIO API errors](https://docs.zettabyte.com/docs/maas/errors)

## Use a key with zCLOUD

Pass the key in the `Authorization` header as `Bearer YOUR_API_KEY` on every inference call. See [Inference API](https://docs.zettabyte.com/docs/inference-api/inference-api).

## Use a key with zSTUDIO

The same header authenticates a call to zSTUDIO's model API: see the [zSTUDIO API overview](https://docs.zettabyte.com/docs/maas/overview) for how a call authenticates and where it goes, [the quickstart](https://docs.zettabyte.com/docs/maas/quickstart) for a first call from code, or [Connect an app or tool to zSTUDIO](https://docs.zettabyte.com/docs/maas/connect-an-app-or-tool) for a settings screen.