Kgateway · Schema

kgateway AIBackend

AIBackend configures AI provider backends with support for LLM providers such as OpenAI, Azure OpenAI, Gemini, and Mistral, including priority-based routing groups.

Gateways

Properties

Name	Type	Description
apiVersion	string	API version for the AIBackend resource.
kind	string	Resource kind.
metadata	object	Standard Kubernetes object metadata.
spec	object	AIBackendSpec defines the desired state of an AIBackend.

View JSON Schema on GitHub

JSON Schema

{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "$id": "https://github.com/api-evangelist/kgateway/blob/main/json-schema/ai-backend.json",
  "title": "kgateway AIBackend",
  "description": "AIBackend configures AI provider backends with support for LLM providers such as OpenAI, Azure OpenAI, Gemini, and Mistral, including priority-based routing groups.",
  "type": "object",
  "properties": {
    "apiVersion": {
      "type": "string",
      "const": "gateway.kgateway.dev/v1alpha1",
      "description": "API version for the AIBackend resource."
    },
    "kind": {
      "type": "string",
      "const": "AIBackend",
      "description": "Resource kind."
    },
    "metadata": {
      "type": "object",
      "description": "Standard Kubernetes object metadata.",
      "properties": {
        "name": {
          "type": "string",
          "description": "Name of the AIBackend resource."
        },
        "namespace": {
          "type": "string",
          "description": "Namespace of the AIBackend resource."
        },
        "labels": {
          "type": "object",
          "additionalProperties": { "type": "string" }
        },
        "annotations": {
          "type": "object",
          "additionalProperties": { "type": "string" }
        }
      },
      "required": ["name"]
    },
    "spec": {
      "type": "object",
      "description": "AIBackendSpec defines the desired state of an AIBackend.",
      "properties": {
        "provider": {
          "type": "object",
          "description": "Configuration for the LLM provider.",
          "properties": {
            "type": {
              "type": "string",
              "description": "The type of LLM provider.",
              "enum": ["openai", "azure", "gemini", "mistral"]
            },
            "openai": {
              "type": "object",
              "properties": {
                "model": {
                  "type": "string",
                  "description": "Model name to use."
                },
                "authToken": { "$ref": "#/$defs/AuthToken" }
              }
            },
            "azure": {
              "type": "object",
              "properties": {
                "endpoint": { "type": "string" },
                "deploymentName": { "type": "string" },
                "apiVersion": { "type": "string" },
                "authToken": { "$ref": "#/$defs/AuthToken" }
              }
            },
            "gemini": {
              "type": "object",
              "properties": {
                "model": { "type": "string" },
                "authToken": { "$ref": "#/$defs/AuthToken" }
              }
            }
          }
        },
        "groups": {
          "type": "array",
          "description": "List of groups in priority order where each group defines a set of LLM providers.",
          "items": {
            "type": "object",
            "properties": {
              "priority": {
                "type": "integer",
                "description": "Priority of this group (lower is higher priority)."
              },
              "backends": {
                "type": "array",
                "items": {
                  "type": "object",
                  "properties": {
                    "backendRef": {
                      "type": "object",
                      "properties": {
                        "name": { "type": "string" },
                        "namespace": { "type": "string" }
                      }
                    },
                    "weight": {
                      "type": "integer",
                      "description": "Weight for load balancing within the group."
                    }
                  }
                }
              }
            }
          }
        }
      }
    }
  },
  "$defs": {
    "AuthToken": {
      "type": "object",
      "properties": {
        "secretRef": {
          "type": "object",
          "properties": {
            "name": { "type": "string" },
            "namespace": { "type": "string" }
          }
        }
      }
    }
  }
}