Authentication

For most normal console use, you do not need to think about authentication at all beyond signing in to RemoteGPU.

The main question for customers is usually simpler:

• when is the console enough?
• when do you need an API key?

Open in the console

Task	Console page
Create or manage API keys	Settings / API Keys

Choose the authentication path

Product path	Typical user	Interface	Do you need an API key?
Application	Non-professional or low-ops user	Console	Usually no
Inference API	Developer or backend integrator	HTTP API	Yes
Kubernetes	Kubernetes-literate operator	Console and kubectl	Required for kubectl; not required for normal console use

When the console is enough

If you are only using the RemoteGPU console to:

• manage Comfy workspaces
• create or review Kubernetes namespaces
• review Kubernetes deployments, services, and ingresses
• manage account settings

you usually do not need an API key.

When you need an API key

You need an API key when you are:

• calling the Inference API
• running kubectl with a RemoteGPU namespace kubeconfig

API keys are created in Settings / API Keys in the console.

Which key should you create

Scope	Use it when	Typical caller
Inference	You need to call inference endpoints only	App server, script, backend worker
Kubernetes exec	You need namespace access through kubectl	Local shell, CI job, platform operator
All	One automation needs both inference and Kubernetes access	Admin automation

Read API keys for scope details and examples.

Typical workflows

Console-only workflow

If you only use the console, sign in with your normal account. No API key is required.

Inference API workflow

1. Create an inference-enabled API key.
2. Pass it in the x-api-key header.
3. Continue with Image inference.

This path is usually the right fit for semi-professional users, developers, and integrators who want a programmable API without operating Kubernetes directly.

Kubernetes workflow

1. Create a Kubernetes exec API key.
2. Download the namespace kubeconfig from the console.
3. Export REMOTEGPU_API_KEY before running kubectl.
4. Continue with Kubernetes overview.

This path is designed for professional users who are comfortable with native Kubernetes resources such as Deployment, Service, and Ingress.

Read next

• Read API keys to choose the right scope.
• Read Image inference if you want to call the inference API.
• Read Kubernetes overview if you want namespace access and kubectl.