1. Docs
  2. Tutorials
  3. Access management
  4. Authentication
  5. As a user

Last modified July 1, 2025

Authentication for the platform API as a user

As a user of the platform API for any given installation, you need:

  • A user account in the identity provider used by the installation (single sign-on).
  • The platform API endpoint of the installation. Alternatively, the web user interface address.

For command-line access to the cluster using kubectl, authentication via our kubectl plugin is required. Detailed instructions are provided below. If you prefer a graphical interface, we also outline the Web UI login method.

Using kubectl gs login

Assuming that you want to work with the API using kubectl, we recommend you install the Giant Swarm plug-in named gs. It can be installed and updated using krew.

To set up your kubectl configuration with a context for your Giant Swarm installation’s platform API, execute the following command:

kubectl gs login URL

As a URL you use the platform API endpoint address (normally starting with https://g8s.).

More information can be found in the kubectl gs login manual page.

Authentication flow

In order to complete the authentication flow, kubectl gs will open your default web browser. The page in the browser will prompt you to select an identity provider you want to use for the authentication. If you are authenticated with the identity provider you select (determined for example via a cookie), you will see a confirmation page and can close the browser window again. However, if you weren’t authenticated with your identity provider yet, you’ll have to go through the authentication process you are used to. When this is done, a confirmation page will be shown.

You can also execute the login command with the --connector-id flag, and specify an identifier of the desired connector (related to your preferred identity provider) for the authentication:

kubectl gs login URL --connector-id CONNECTOR_ID

As a result of running the command, your kubectl configuration has a new context, user, and cluster entry. The context is named according to the pattern

gs-INSTALLATION_NAME

INSTALLATION_NAME is the unique name of your Giant Swarm installation.

This context is selected automatically as the current context, so you are ready to use kubectl with the platform API.

When switching back to this context, it shouldn’t be necessary to go through the web-based authentication flow again. kubectl will automatically refresh the authentication token when needed, without your interaction.

Further reading