Create and manage AI Gateway API keys to authenticate application requests, and optionally attach a routing configuration so every request using that key.
| Where Can I Use This? | What Do I Need? |
- Prisma AIRS AI Gateway (Americas region)
|
- AI Gateway activated
- At least one LLM integration configured
|
AI Gateway uses two types of credentials to authenticate requests: the gateway API key
and the virtual key. The gateway API key (sent in the
x-portkey-api-key header) authenticates the application or service making
the request to AI Gateway. The virtual key (sent in the
x-portkey-virtual-key header) identifies the workspace and grants access
to that workspace's LLM integrations. Both headers are required for most requests.
You can optionally attach a routing
configuration to an API key. When a config is attached,
every request that uses that API key automatically applies the config's load balancing,
fallback, caching, or guardrail settings, regardless of whether the request body specifies
a config explicitly. This lets you enforce a standard routing policy for an application
without requiring the developer to reference a config in every call.
Gateway API keys are scoped to one of three roles: organization service
(ORGANISATION_SERVICE), workspace user (WORKSPACE_USER), or workspace
service (WORKSPACE_SERVICE). Organization service keys have the broadest scope;
workspace-scoped keys are limited to a single workspace.