@databricks/sdk-modelserving 0.0.0-dev → 0.1.0-dev.2

package/src/v1/model.ts

@@ -0,0 +1,2589 @@
+// Code generated from API definition by Databricks SDK Generator. DO NOT EDIT.
+
+import {z} from 'zod';
+
+export enum Behavior {
+  BEHAVIOR_UNSPECIFIED = 'BEHAVIOR_UNSPECIFIED',
+  NONE = 'NONE',
+  BLOCK = 'BLOCK',
+  MASK = 'MASK',
+}
+
+export enum ServedModelDeploymentState {
+  DEPLOYMENT_UNKNOWN = 'DEPLOYMENT_UNKNOWN',
+  DEPLOYMENT_CREATING = 'DEPLOYMENT_CREATING',
+  DEPLOYMENT_RECOVERING = 'DEPLOYMENT_RECOVERING',
+  DEPLOYMENT_READY = 'DEPLOYMENT_READY',
+  DEPLOYMENT_FAILED = 'DEPLOYMENT_FAILED',
+  DEPLOYMENT_ABORTED = 'DEPLOYMENT_ABORTED',
+  DEPLOYMENT_STOPPED = 'DEPLOYMENT_STOPPED',
+}
+
+export enum ServingEndpointDetailedPermissionLevel {
+  CAN_MANAGE = 'CAN_MANAGE',
+  CAN_QUERY = 'CAN_QUERY',
+  CAN_VIEW = 'CAN_VIEW',
+}
+
+// eslint-disable-next-line @typescript-eslint/naming-convention -- Proto-style nested enum name.
+export enum ExternalFunctionRequest_HttpMethod {
+  HTTP_METHOD_UNSPECIFIED = 'HTTP_METHOD_UNSPECIFIED',
+  GET = 'GET',
+  POST = 'POST',
+  PUT = 'PUT',
+  DELETE = 'DELETE',
+  PATCH = 'PATCH',
+}
+
+// eslint-disable-next-line @typescript-eslint/naming-convention -- Proto-style nested enum name.
+export enum InferenceEndpointState_ConfigUpdateState {
+  CONFIG_UPDATE_STATE_UNSPECIFIED = 'CONFIG_UPDATE_STATE_UNSPECIFIED',
+  NOT_UPDATING = 'NOT_UPDATING',
+  IN_PROGRESS = 'IN_PROGRESS',
+  UPDATE_FAILED = 'UPDATE_FAILED',
+  UPDATE_CANCELED = 'UPDATE_CANCELED',
+}
+
+// eslint-disable-next-line @typescript-eslint/naming-convention -- Proto-style nested enum name.
+export enum InferenceEndpointState_ReadyState {
+  READY_STATE_UNSPECIFIED = 'READY_STATE_UNSPECIFIED',
+  READY = 'READY',
+  NOT_READY = 'NOT_READY',
+}
+
+export interface Ai21LabsConfig {
+  /**
+   * The <Databricks> secret key reference for an AI21 Labs API key. If you
+   * prefer to paste your API key directly, see `ai21labs_api_key_plaintext`.
+   * You must provide an API key using one of the following fields:
+   * `ai21labs_api_key` or `ai21labs_api_key_plaintext`.
+   */
+  ai21labsApiKey?: string | undefined;
+  /**
+   * An AI21 Labs API key provided as a plaintext string. If you prefer to
+   * reference your key using Databricks Secrets, see `ai21labs_api_key`. You
+   * must provide an API key using one of the following fields:
+   * `ai21labs_api_key` or `ai21labs_api_key_plaintext`.
+   */
+  ai21labsApiKeyPlaintext?: string | undefined;
+}
+
+export interface AiGatewayConfig {
+  /**
+   * Configuration to enable usage tracking using system tables.
+   * These tables allow you to monitor operational usage on endpoints and their associated costs.
+   */
+  usageTrackingConfig?: UsageTrackingConfig | undefined;
+  /**
+   * Configuration for payload logging using inference tables.
+   * Use these tables to monitor and audit data being sent to and received from model APIs and to improve model quality.
+   */
+  inferenceTableConfig?: InferenceTableConfig | undefined;
+  /** Configuration for rate limits which can be set to limit endpoint traffic. */
+  rateLimits?: AiGatewayRateLimit[] | undefined;
+  /** Configuration for AI Guardrails to prevent unwanted data and unsafe data in requests and responses. */
+  guardrails?: AiGuardrails | undefined;
+  /**
+   * Configuration for traffic fallback which auto fallbacks to other served entities if the request to a served
+   * entity fails with certain error codes, to increase availability.
+   */
+  fallbackConfig?: FallbackConfig | undefined;
+}
+
+export interface AiGatewayRateLimit {
+  /** Used to specify how many calls are allowed for a key within the renewal_period. */
+  calls?: bigint | undefined;
+  /**
+   * Key field for a rate limit. Currently, 'user', 'user_group, 'service_principal', and 'endpoint' are supported,
+   * with 'endpoint' being the default if not specified.
+   */
+  key?: string | undefined;
+  /** Renewal period field for a rate limit. Currently, only 'minute' is supported. */
+  renewalPeriod?: string | undefined;
+  /** Principal field for a user, user group, or service principal to apply rate limiting to. Accepts a user email, group name, or service principal application ID. */
+  principal?: string | undefined;
+  /** Used to specify how many tokens are allowed for a key within the renewal_period. */
+  tokens?: bigint | undefined;
+}
+
+export interface AiGuardrailParameters {
+  /** Indicates whether the safety filter is enabled. */
+  safety?: boolean | undefined;
+  /** Configuration for guardrail PII filter. */
+  pii?: PiiSettings | undefined;
+  /**
+   * The list of allowed topics.
+   * Given a chat request, this guardrail flags the request if its topic is not in the allowed topics.
+   */
+  validTopics?: string[] | undefined;
+  /**
+   * List of invalid keywords.
+   * AI guardrail uses keyword or string matching to decide if the keyword exists in the request or response content.
+   */
+  invalidKeywords?: string[] | undefined;
+}
+
+export interface AiGuardrails {
+  /** Configuration for input guardrail filters. */
+  input?: AiGuardrailParameters | undefined;
+  /** Configuration for output guardrail filters. */
+  output?: AiGuardrailParameters | undefined;
+}
+
+export interface AmazonBedrockConfig {
+  /** The AWS region to use. Bedrock has to be enabled there. */
+  awsRegion?: string | undefined;
+  /**
+   * The <Databricks> secret key reference for an AWS access key ID with
+   * permissions to interact with Bedrock services. If you prefer to paste
+   * your API key directly, see `aws_access_key_id_plaintext`. You must provide an API
+   * key using one of the following fields: `aws_access_key_id` or
+   * `aws_access_key_id_plaintext`.
+   */
+  awsAccessKeyId?: string | undefined;
+  /**
+   * The <Databricks> secret key reference for an AWS secret access key paired
+   * with the access key ID, with permissions to interact with Bedrock
+   * services. If you prefer to paste your API key directly, see
+   * `aws_secret_access_key_plaintext`. You must provide an API key using one
+   * of the following fields: `aws_secret_access_key` or
+   * `aws_secret_access_key_plaintext`.
+   */
+  awsSecretAccessKey?: string | undefined;
+  /**
+   * The underlying provider in Amazon Bedrock. Supported values (case
+   * insensitive) include: Anthropic, Cohere, AI21Labs, Amazon.
+   */
+  bedrockProvider?: string | undefined;
+  /**
+   * An AWS access key ID with permissions to interact with Bedrock services
+   * provided as a plaintext string. If you prefer to reference your key using
+   * Databricks Secrets, see `aws_access_key_id`. You must provide an API key
+   * using one of the following fields: `aws_access_key_id` or
+   * `aws_access_key_id_plaintext`.
+   */
+  awsAccessKeyIdPlaintext?: string | undefined;
+  /**
+   * An AWS secret access key paired with the access key ID, with permissions
+   * to interact with Bedrock services provided as a plaintext string. If you
+   * prefer to reference your key using Databricks Secrets, see
+   * `aws_secret_access_key`. You must provide an API key using one of the
+   * following fields: `aws_secret_access_key` or
+   * `aws_secret_access_key_plaintext`.
+   */
+  awsSecretAccessKeyPlaintext?: string | undefined;
+  /**
+   * ARN of the instance profile that the external model will use to access AWS resources.
+   * You must authenticate using an instance profile or access keys.
+   * If you prefer to authenticate using access keys, see `aws_access_key_id`,
+   * `aws_access_key_id_plaintext`, `aws_secret_access_key` and `aws_secret_access_key_plaintext`.
+   */
+  instanceProfileArn?: string | undefined;
+}
+
+export interface AnthropicConfig {
+  /**
+   * The <Databricks> secret key reference for an Anthropic API key. If you
+   * prefer to paste your API key directly, see `anthropic_api_key_plaintext`.
+   * You must provide an API key using one of the following fields:
+   * `anthropic_api_key` or `anthropic_api_key_plaintext`.
+   */
+  anthropicApiKey?: string | undefined;
+  /**
+   * The Anthropic API key provided as a plaintext string. If you prefer to
+   * reference your key using Databricks Secrets, see `anthropic_api_key`. You
+   * must provide an API key using one of the following fields:
+   * `anthropic_api_key` or `anthropic_api_key_plaintext`.
+   */
+  anthropicApiKeyPlaintext?: string | undefined;
+}
+
+export interface ApiKeyAuth {
+  /** The name of the API key parameter used for authentication. */
+  key?: string | undefined;
+  /**
+   * The <Databricks> secret key reference for an API Key.
+   * If you prefer to paste your token directly, see `value_plaintext`.
+   */
+  value?: string | undefined;
+  /**
+   * The API Key provided as a plaintext string. If you prefer to reference your
+   * token using Databricks Secrets, see `value`.
+   */
+  valuePlaintext?: string | undefined;
+}
+
+/**
+ * Deprecated: legacy inference table configuration. Please use AI Gateway inference tables instead.
+ * See https://docs.databricks.com/aws/en/ai-gateway/inference-tables.
+ */
+export interface AutoCaptureConfig {
+  /** The name of the catalog in Unity Catalog. NOTE: On update, you cannot change the catalog name if the inference table is already enabled. */
+  catalogName?: string | undefined;
+  /** The name of the schema in Unity Catalog. NOTE: On update, you cannot change the schema name if the inference table is already enabled. */
+  schemaName?: string | undefined;
+  /** The prefix of the table in Unity Catalog. NOTE: On update, you cannot change the prefix name if the inference table is already enabled. */
+  tableNamePrefix?: string | undefined;
+  state?: AutoCaptureState | undefined;
+  /** Indicates whether the inference table is enabled. */
+  enabled?: boolean | undefined;
+}
+
+export interface AutoCaptureState {
+  payloadTable?: PayloadTable | undefined;
+}
+
+export interface BearerTokenAuth {
+  /**
+   * The <Databricks> secret key reference for a token.
+   * If you prefer to paste your token directly, see `token_plaintext`.
+   */
+  token?: string | undefined;
+  /**
+   * The token provided as a plaintext string. If you prefer to reference your
+   * token using Databricks Secrets, see `token`.
+   */
+  tokenPlaintext?: string | undefined;
+}
+
+export interface CohereConfig {
+  /**
+   * The <Databricks> secret key reference for a Cohere API key. If you prefer
+   * to paste your API key directly, see `cohere_api_key_plaintext`. You must
+   * provide an API key using one of the following fields: `cohere_api_key` or
+   * `cohere_api_key_plaintext`.
+   */
+  cohereApiKey?: string | undefined;
+  /**
+   * The Cohere API key provided as a plaintext string. If you prefer to
+   * reference your key using Databricks Secrets, see `cohere_api_key`. You
+   * must provide an API key using one of the following fields:
+   * `cohere_api_key` or `cohere_api_key_plaintext`.
+   */
+  cohereApiKeyPlaintext?: string | undefined;
+  /**
+   * This is an optional field to provide a customized base URL for the Cohere
+   * API. If left unspecified, the standard Cohere base URL is used.
+   */
+  cohereApiBase?: string | undefined;
+}
+
+export interface CreateInferenceEndpointRequest {
+  /**
+   * The name of the serving endpoint. This field is required and must be unique across a <Workspace>.
+   * An endpoint name can consist of alphanumeric characters, dashes, and underscores.
+   */
+  name?: string | undefined;
+  /** The core config of the serving endpoint. */
+  config?: EndpointCoreConfig | undefined;
+  /** Tags to be attached to the serving endpoint and automatically propagated to billing logs. */
+  tags?: EndpointTag[] | undefined;
+  /** Enable route optimization for the serving endpoint. */
+  routeOptimized?: boolean | undefined;
+  /** Rate limits to be applied to the serving endpoint. NOTE: this field is deprecated, please use AI Gateway to manage rate limits. */
+  rateLimits?: RateLimit[] | undefined;
+  /** The AI Gateway configuration for the serving endpoint. NOTE: External model, provisioned throughput, and pay-per-token endpoints are fully supported; agent endpoints currently only support inference tables. */
+  aiGateway?: AiGatewayConfig | undefined;
+  /** The budget policy to be applied to the serving endpoint. */
+  budgetPolicyId?: string | undefined;
+  /** Email notification settings. */
+  emailNotifications?: EmailNotifications | undefined;
+  description?: string | undefined;
+}
+
+export interface CreatePtEndpointRequest {
+  /**
+   * The name of the serving endpoint. This field is required and must be unique across a <Workspace>.
+   * An endpoint name can consist of alphanumeric characters, dashes, and underscores.
+   */
+  name?: string | undefined;
+  /** The core config of the serving endpoint. */
+  config?: PtEndpointCoreConfig | undefined;
+  /** Tags to be attached to the serving endpoint and automatically propagated to billing logs. */
+  tags?: EndpointTag[] | undefined;
+  /** The AI Gateway configuration for the serving endpoint. */
+  aiGateway?: AiGatewayConfig | undefined;
+  /** The budget policy associated with the endpoint. */
+  budgetPolicyId?: string | undefined;
+  /** Email notification settings. */
+  emailNotifications?: EmailNotifications | undefined;
+}
+
+/** Configs needed to create a custom provider model route. */
+export interface CustomProviderConfig {
+  /** This is a field to provide the URL of the custom provider API. */
+  customProviderUrl?: string | undefined;
+  /**
+   * This is a field to provide bearer token authentication for the custom provider API.
+   * You can only specify one authentication method.
+   */
+  bearerTokenAuth?: BearerTokenAuth | undefined;
+  /**
+   * This is a field to provide API key authentication for the custom provider API.
+   * You can only specify one authentication method.
+   */
+  apiKeyAuth?: ApiKeyAuth | undefined;
+}
+
+/** Details necessary to query this object's API through the DataPlane APIs. */
+export interface DataPlaneInfo {
+  /** The URL of the endpoint for this operation in the dataplane. */
+  endpointUrl?: string | undefined;
+  /** Authorization details as a string. */
+  authorizationDetails?: string | undefined;
+}
+
+export interface DatabricksModelServingConfig {
+  /**
+   * The <Databricks> secret key reference for a Databricks API token that
+   * corresponds to a user or service principal with Can Query access to the
+   * model serving endpoint pointed to by this external model. If you prefer
+   * to paste your API key directly, see `databricks_api_token_plaintext`. You
+   * must provide an API key using one of the following fields:
+   * `databricks_api_token` or `databricks_api_token_plaintext`.
+   */
+  databricksApiToken?: string | undefined;
+  /**
+   * The URL of the <Databricks> workspace containing the model serving endpoint
+   * pointed to by this external model.
+   */
+  databricksWorkspaceUrl?: string | undefined;
+  /**
+   * The Databricks API token that corresponds to a user or service principal
+   * with Can Query access to the model serving endpoint pointed to by this
+   * external model provided as a plaintext string. If you prefer to reference
+   * your key using Databricks Secrets, see `databricks_api_token`. You must
+   * provide an API key using one of the following fields:
+   * `databricks_api_token` or `databricks_api_token_plaintext`.
+   */
+  databricksApiTokenPlaintext?: string | undefined;
+}
+
+export interface DeleteInferenceEndpointRequest {
+  name?: string | undefined;
+}
+
+// eslint-disable-next-line @typescript-eslint/naming-convention, @typescript-eslint/no-empty-object-type -- Proto-style nested message name.
+export interface DeleteInferenceEndpointRequest_Response {}
+
+export interface EmailNotifications {
+  /** A list of email addresses to be notified when an endpoint successfully updates its configuration or state. */
+  onUpdateSuccess?: string[] | undefined;
+  /** A list of email addresses to be notified when an endpoint fails to update its configuration or state. */
+  onUpdateFailure?: string[] | undefined;
+}
+
+export interface EndpointCoreConfig {
+  /** The list of served entities under the serving endpoint config. */
+  servedEntities?: ServedModel[] | undefined;
+  /** (Deprecated, use served_entities instead) The list of served models under the serving endpoint config. */
+  servedModels?: ServedModel[] | undefined;
+  /** The traffic configuration associated with the serving endpoint config. */
+  trafficConfig?: TrafficConfig | undefined;
+  /**
+   * Configuration for legacy Inference Tables which automatically log requests and responses to Unity
+   * Catalog.
+   * Deprecated: please use AI Gateway inference tables instead. See
+   * https://docs.databricks.com/aws/en/ai-gateway/inference-tables.
+   */
+  autoCaptureConfig?: AutoCaptureConfig | undefined;
+}
+
+export interface EndpointCoreConfigOutput {
+  /** The config version that the serving endpoint is currently serving. */
+  configVersion?: bigint | undefined;
+  /** The list of served entities under the serving endpoint config. */
+  servedEntities?: ServedModel[] | undefined;
+  /** (Deprecated, use served_entities instead) The list of served models under the serving endpoint config. */
+  servedModels?: ServedModel[] | undefined;
+  /** The traffic configuration associated with the serving endpoint config. */
+  trafficConfig?: TrafficConfig | undefined;
+  /**
+   * Configuration for legacy Inference Tables which automatically log requests and responses to Unity
+   * Catalog.
+   * Deprecated: please use AI Gateway inference tables instead. See
+   * https://docs.databricks.com/aws/en/ai-gateway/inference-tables.
+   */
+  autoCaptureConfig?: AutoCaptureConfig | undefined;
+}
+
+export interface EndpointCoreConfigSummary {
+  /** The list of served entities under the serving endpoint config. */
+  servedEntities?: ServedModelLite[] | undefined;
+  /** (Deprecated, use served_entities instead) The list of served models under the serving endpoint config. */
+  servedModels?: ServedModelLite[] | undefined;
+}
+
+export interface EndpointTag {
+  /** Key field for a serving endpoint tag. */
+  key?: string | undefined;
+  /** Optional value field for a serving endpoint tag. */
+  value?: string | undefined;
+}
+
+/**
+ * *
+ * Proto version of com.databricks.rpc.HttpOverRpcResponse.
+ *
+ * This message can be specially handled in UnaryRpcService with JettyRPC when the advanced feature
+ * CustomHandlingForHttpOverRpcProtoResponse is enabled - bypass the RPC serializer and populate
+ * HTTP status, response headers and response body from the proto message directly.
+ *
+ * Don't add/modify the fields before being aware of the implications.
+ */
+export interface ExportMetricsResponse {
+  contents?: ReadableStream | undefined;
+}
+
+/** Simple Proto message for testing */
+export interface ExternalFunctionRequest {
+  /** The connection name to use. This is required to identify the external connection. */
+  connectionName?: string | undefined;
+  /** The HTTP method to use (e.g., 'GET', 'POST'). */
+  method?: ExternalFunctionRequest_HttpMethod | undefined;
+  /** The relative path for the API endpoint. This is required. */
+  path?: string | undefined;
+  /** The JSON payload to send in the request body. */
+  json?: string | undefined;
+  /** Additional headers for the request. If not provided, only auth headers from connections would be passed. */
+  headers?: string | undefined;
+  /** Query parameters for the request. */
+  params?: string | undefined;
+  /**
+   * Optional subdomain to prepend to the connection URL's host. If provided, this will be
+   * added as a prefix to the connection URL's host. For example, if the connection URL is
+   * `https://api.example.com/v1` and `sub_domain` is `"custom"`, the resulting URL will be
+   * `https://custom.api.example.com/v1`.
+   */
+  subDomain?: string | undefined;
+}
+
+export interface ExternalFunctionResponse {
+  contents?: ReadableStream | undefined;
+}
+
+export interface ExternalModel {
+  /** The name of the provider for the external model. Currently, the supported providers are 'ai21labs', 'anthropic', 'amazon-bedrock', 'cohere', 'databricks-model-serving', 'google-cloud-vertex-ai', 'openai', 'palm', and 'custom'. */
+  provider?: string | undefined;
+  /** The name of the external model. */
+  name?: string | undefined;
+  /** The task type of the external model. */
+  task?: string | undefined;
+  /** external model config. The config corresponding to the provider will be used. */
+  config?:
+    | {
+        $case: 'ai21labsConfig';
+        /** AI21Labs Config. Only required if the provider is 'ai21labs'. */
+        ai21labsConfig: Ai21LabsConfig;
+      }
+    | {
+        $case: 'anthropicConfig';
+        /** Anthropic Config. Only required if the provider is 'anthropic'. */
+        anthropicConfig: AnthropicConfig;
+      }
+    | {
+        $case: 'amazonBedrockConfig';
+        /** Amazon Bedrock Config. Only required if the provider is 'amazon-bedrock'. */
+        amazonBedrockConfig: AmazonBedrockConfig;
+      }
+    | {
+        $case: 'cohereConfig';
+        /** Cohere Config. Only required if the provider is 'cohere'. */
+        cohereConfig: CohereConfig;
+      }
+    | {
+        $case: 'googleCloudVertexAiConfig';
+        /** Google Cloud Vertex AI Config. Only required if the provider is 'google-cloud-vertex-ai'. */
+        googleCloudVertexAiConfig: GoogleCloudVertexAiConfig;
+      }
+    | {
+        $case: 'databricksModelServingConfig';
+        /** Databricks Model Serving Config. Only required if the provider is 'databricks-model-serving'. */
+        databricksModelServingConfig: DatabricksModelServingConfig;
+      }
+    | {
+        $case: 'openaiConfig';
+        /** OpenAI Config. Only required if the provider is 'openai'. */
+        openaiConfig: OpenAiConfig;
+      }
+    | {
+        $case: 'palmConfig';
+        /** PaLM Config. Only required if the provider is 'palm'. */
+        palmConfig: PaLmConfig;
+      }
+    | {
+        $case: 'customProviderConfig';
+        /** Custom Provider Config. Only required if the provider is 'custom'. */
+        customProviderConfig: CustomProviderConfig;
+      }
+    | undefined;
+}
+
+export interface FallbackConfig {
+  /**
+   * Whether to enable traffic fallback. When a served entity in the serving endpoint returns specific error
+   * codes (e.g. 500), the request will automatically be round-robin attempted with other served entities in the same
+   * endpoint, following the order of served entity list, until a successful response is returned.
+   * If all attempts fail, return the last response with the error code.
+   */
+  enabled?: boolean | undefined;
+}
+
+/** All fields are not sensitive as they are hard-coded in the system and made available to customers. */
+export interface FoundationModel {
+  name?: string | undefined;
+  displayName?: string | undefined;
+  docs?: string | undefined;
+  description?: string | undefined;
+}
+
+export interface GetExportEndpointMetricsRequest {
+  /** The name of the serving endpoint to retrieve metrics for. This field is required. */
+  name?: string | undefined;
+}
+
+export interface GetInferenceEndpointRequest {
+  /** The name of the serving endpoint. This field is required. */
+  name?: string | undefined;
+}
+
+export interface GetInferenceEndpointSchemaRequest {
+  /** The name of the serving endpoint that the served model belongs to. This field is required. */
+  name?: string | undefined;
+}
+
+/** The top level proto message that represents an OpenAPI 3.0 document. */
+export interface GetOpenApiResponse {
+  contents?: ReadableStream | undefined;
+}
+
+export interface GetServedModelBuildLogsRequest {
+  /** The name of the serving endpoint that the served model belongs to. This field is required. */
+  name?: string | undefined;
+  /** The name of the served model that build logs will be retrieved for. This field is required. */
+  servedModelName?: string | undefined;
+}
+
+// eslint-disable-next-line @typescript-eslint/naming-convention -- Proto-style nested message name.
+export interface GetServedModelBuildLogsRequest_Response {
+  /** The logs associated with building the served entity's environment. */
+  logs?: string | undefined;
+}
+
+export interface GetServedModelLogsRequest {
+  /** The name of the serving endpoint that the served model belongs to. This field is required. */
+  name?: string | undefined;
+  /** The name of the served model that logs will be retrieved for. This field is required. */
+  servedModelName?: string | undefined;
+}
+
+// eslint-disable-next-line @typescript-eslint/naming-convention -- Proto-style nested message name.
+export interface GetServedModelLogsRequest_Response {
+  /** The most recent log lines of the model server processing invocation requests. */
+  logs?: string | undefined;
+}
+
+export interface GoogleCloudVertexAiConfig {
+  /**
+   * The <Databricks> secret key reference for a private key for the service
+   * account which has access to the Google Cloud Vertex AI Service. See [Best
+   * practices for managing service account keys]. If you prefer to paste your
+   * API key directly, see `private_key_plaintext`. You must provide an API
+   * key using one of the following fields: `private_key` or
+   * `private_key_plaintext`
+   *
+   * [Best practices for managing service account keys]: https://cloud.google.com/iam/docs/best-practices-for-managing-service-account-keys
+   */
+  privateKey?: string | undefined;
+  /**
+   * This is the Google Cloud project id that the service account is
+   * associated with.
+   */
+  projectId?: string | undefined;
+  /**
+   * This is the region for the Google Cloud Vertex AI Service. See [supported
+   * regions] for more details. Some models are only available in specific
+   * regions.
+   *
+   * [supported regions]: https://cloud.google.com/vertex-ai/docs/general/locations
+   */
+  region?: string | undefined;
+  /**
+   * The private key for the service account which has access to the Google
+   * Cloud Vertex AI Service provided as a plaintext secret. See [Best
+   * practices for managing service account keys]. If you prefer to reference
+   * your key using Databricks Secrets, see `private_key`. You must provide an
+   * API key using one of the following fields: `private_key` or
+   * `private_key_plaintext`.
+   *
+   * [Best practices for managing service account keys]: https://cloud.google.com/iam/docs/best-practices-for-managing-service-account-keys
+   */
+  privateKeyPlaintext?: string | undefined;
+}
+
+export interface InferenceEndpoint {
+  /** The name of the serving endpoint. */
+  name?: string | undefined;
+  /** The email of the user who created the serving endpoint. */
+  creator?: string | undefined;
+  /** The timestamp when the endpoint was created in Unix time. */
+  creationTimestamp?: bigint | undefined;
+  /** The timestamp when the endpoint was last updated by a user in Unix time. */
+  lastUpdatedTimestamp?: bigint | undefined;
+  /** Information corresponding to the state of the serving endpoint. */
+  state?: InferenceEndpointState | undefined;
+  /** The config that is currently being served by the endpoint. */
+  config?: EndpointCoreConfigSummary | undefined;
+  /** Tags attached to the serving endpoint. */
+  tags?: EndpointTag[] | undefined;
+  /** System-generated ID of the endpoint, included to be used by the Permissions API. */
+  id?: string | undefined;
+  /** The task type of the serving endpoint. */
+  task?: string | undefined;
+  /** The AI Gateway configuration for the serving endpoint. NOTE: External model, provisioned throughput, and pay-per-token endpoints are fully supported; agent endpoints currently only support inference tables. */
+  aiGateway?: AiGatewayConfig | undefined;
+  /** The budget policy associated with the endpoint. */
+  budgetPolicyId?: string | undefined;
+  /** Description of the endpoint */
+  description?: string | undefined;
+  /** The usage policy associated with serving endpoint. */
+  usagePolicyId?: string | undefined;
+}
+
+export interface InferenceEndpointDetailed {
+  /** The name of the serving endpoint. */
+  name?: string | undefined;
+  /** The email of the user who created the serving endpoint. */
+  creator?: string | undefined;
+  /** The timestamp when the endpoint was created in Unix time. */
+  creationTimestamp?: bigint | undefined;
+  /** The timestamp when the endpoint was last updated by a user in Unix time. */
+  lastUpdatedTimestamp?: bigint | undefined;
+  /** Information corresponding to the state of the serving endpoint. */
+  state?: InferenceEndpointState | undefined;
+  /** The config that is currently being served by the endpoint. */
+  config?: EndpointCoreConfigOutput | undefined;
+  /** The config that the endpoint is attempting to update to. */
+  pendingConfig?: PendingConfig | undefined;
+  /** System-generated ID of the endpoint. This is used to refer to the endpoint in the Permissions API */
+  id?: string | undefined;
+  /** The permission level of the principal making the request. */
+  permissionLevel?: ServingEndpointDetailedPermissionLevel | undefined;
+  /** Tags attached to the serving endpoint. */
+  tags?: EndpointTag[] | undefined;
+  /** The task type of the serving endpoint. */
+  task?: string | undefined;
+  /** Boolean representing if route optimization has been enabled for the endpoint */
+  routeOptimized?: boolean | undefined;
+  /** Endpoint invocation url if route optimization is enabled for endpoint */
+  endpointUrl?: string | undefined;
+  /** Information required to query DataPlane APIs. */
+  dataPlaneInfo?: ModelDataPlaneInfo | undefined;
+  /** The AI Gateway configuration for the serving endpoint. NOTE: External model, provisioned throughput, and pay-per-token endpoints are fully supported; agent endpoints currently only support inference tables. */
+  aiGateway?: AiGatewayConfig | undefined;
+  /** The budget policy associated with the endpoint. */
+  budgetPolicyId?: string | undefined;
+  /** Email notification settings. */
+  emailNotifications?: EmailNotifications | undefined;
+  /** Description of the serving model */
+  description?: string | undefined;
+}
+
+export interface InferenceEndpointState {
+  /**
+   * The state of an endpoint, indicating whether or not the endpoint is queryable. An endpoint is READY if all of
+   * the served entities in its active configuration are ready. If any of the actively served entities are in a
+   * non-ready state, the endpoint state will be NOT_READY.
+   */
+  ready?: InferenceEndpointState_ReadyState | undefined;
+  /**
+   * The state of an endpoint's config update. This informs the user if the pending_config is in progress, if the
+   * update failed, or if there is no update in progress. Note that if the endpoint's config_update state value is
+   * IN_PROGRESS, another update can not be made until the update completes or fails.
+   */
+  configUpdate?: InferenceEndpointState_ConfigUpdateState | undefined;
+}
+
+export interface InferenceTableConfig {
+  /**
+   * The name of the catalog in Unity Catalog. Required when enabling inference tables.
+   * NOTE: On update, you have to disable inference table first in order to change the catalog name.
+   */
+  catalogName?: string | undefined;
+  /**
+   * The name of the schema in Unity Catalog. Required when enabling inference tables.
+   * NOTE: On update, you have to disable inference table first in order to change the schema name.
+   */
+  schemaName?: string | undefined;
+  /**
+   * The prefix of the table in Unity Catalog.
+   * NOTE: On update, you have to disable inference table first in order to change the prefix name.
+   */
+  tableNamePrefix?: string | undefined;
+  /** Indicates whether the inference table is enabled. */
+  enabled?: boolean | undefined;
+}
+
+// eslint-disable-next-line @typescript-eslint/no-empty-object-type
+export interface ListInferenceEndpointsRequest {}
+
+// eslint-disable-next-line @typescript-eslint/naming-convention -- Proto-style nested message name.
+export interface ListInferenceEndpointsRequest_Response {
+  /** The list of endpoints. */
+  endpoints?: InferenceEndpoint[] | undefined;
+}
+
+/** A representation of all DataPlaneInfo for operations that can be done on a model through Data Plane APIs. */
+export interface ModelDataPlaneInfo {
+  /** Information required to query DataPlane API 'query' endpoint. */
+  queryInfo?: DataPlaneInfo | undefined;
+}
+
+/** Configs needed to create an OpenAI model route. */
+export interface OpenAiConfig {
+  /**
+   * The <Databricks> secret key reference for an OpenAI API key using the
+   * OpenAI or Azure service. If you prefer to paste your API key directly,
+   * see `openai_api_key_plaintext`. You must provide an API key using one of
+   * the following fields: `openai_api_key` or `openai_api_key_plaintext`.
+   */
+  openaiApiKey?: string | undefined;
+  /**
+   * This is an optional field to specify the type of OpenAI API to use. For
+   * Azure OpenAI, this field is required, and adjust this parameter to
+   * represent the preferred security access validation protocol. For access
+   * token validation, use azure. For authentication using Azure Active
+   * Directory (Azure AD) use, azuread.
+   */
+  openaiApiType?: string | undefined;
+  /**
+   * This is a field to provide a customized base URl for the OpenAI API. For
+   * Azure OpenAI, this field is required, and is the base URL for the Azure
+   * OpenAI API service provided by Azure. For other OpenAI API types, this
+   * field is optional, and if left unspecified, the standard OpenAI base URL
+   * is used.
+   */
+  openaiApiBase?: string | undefined;
+  /**
+   * This is an optional field to specify the OpenAI API version. For Azure
+   * OpenAI, this field is required, and is the version of the Azure OpenAI
+   * service to utilize, specified by a date.
+   */
+  openaiApiVersion?: string | undefined;
+  /**
+   * This field is only required for Azure OpenAI and is the name of the
+   * deployment resource for the Azure OpenAI service.
+   */
+  openaiDeploymentName?: string | undefined;
+  /**
+   * This is an optional field to specify the organization in OpenAI or Azure
+   * OpenAI.
+   */
+  openaiOrganization?: string | undefined;
+  /**
+   * This field is only required for Azure AD OpenAI and is the Microsoft
+   * Entra Tenant ID.
+   */
+  microsoftEntraTenantId?: string | undefined;
+  /**
+   * This field is only required for Azure AD OpenAI and is the Microsoft
+   * Entra Client ID.
+   */
+  microsoftEntraClientId?: string | undefined;
+  /**
+   * The <Databricks> secret key reference for a client secret used for
+   * Microsoft Entra ID authentication. If you prefer to paste your client
+   * secret directly, see `microsoft_entra_client_secret_plaintext`. You must
+   * provide an API key using one of the following fields:
+   * `microsoft_entra_client_secret` or
+   * `microsoft_entra_client_secret_plaintext`.
+   */
+  microsoftEntraClientSecret?: string | undefined;
+  /**
+   * The OpenAI API key using the OpenAI or Azure service provided as a
+   * plaintext string. If you prefer to reference your key using Databricks
+   * Secrets, see `openai_api_key`. You must provide an API key using one of
+   * the following fields: `openai_api_key` or `openai_api_key_plaintext`.
+   */
+  openaiApiKeyPlaintext?: string | undefined;
+  /**
+   * The client secret used for Microsoft Entra ID authentication provided as
+   * a plaintext string. If you prefer to reference your key using Databricks
+   * Secrets, see `microsoft_entra_client_secret`. You must provide an API key
+   * using one of the following fields: `microsoft_entra_client_secret` or
+   * `microsoft_entra_client_secret_plaintext`.
+   */
+  microsoftEntraClientSecretPlaintext?: string | undefined;
+}
+
+export interface PaLmConfig {
+  /**
+   * The <Databricks> secret key reference for a PaLM API key. If you prefer to
+   * paste your API key directly, see `palm_api_key_plaintext`. You must
+   * provide an API key using one of the following fields: `palm_api_key` or
+   * `palm_api_key_plaintext`.
+   */
+  palmApiKey?: string | undefined;
+  /**
+   * The PaLM API key provided as a plaintext string. If you prefer to
+   * reference your key using Databricks Secrets, see `palm_api_key`. You must
+   * provide an API key using one of the following fields: `palm_api_key` or
+   * `palm_api_key_plaintext`.
+   */
+  palmApiKeyPlaintext?: string | undefined;
+}
+
+export interface PatchInferenceEndpointTagsRequest {
+  /** The name of the serving endpoint who's tags to patch. This field is required. */
+  name?: string | undefined;
+  /** List of endpoint tags to add */
+  addTags?: EndpointTag[] | undefined;
+  /** List of tag keys to delete */
+  deleteTags?: string[] | undefined;
+}
+
+// eslint-disable-next-line @typescript-eslint/naming-convention -- Proto-style nested message name.
+export interface PatchInferenceEndpointTagsRequest_Response {
+  tags?: EndpointTag[] | undefined;
+}
+
+export interface PayloadTable {
+  name?: string | undefined;
+  status?: string | undefined;
+  statusMessage?: string | undefined;
+}
+
+export interface PendingConfig {
+  /** The list of served entities belonging to the last issued update to the serving endpoint. */
+  servedEntities?: ServedModel[] | undefined;
+  /** (Deprecated, use served_entities instead) The list of served models belonging to the last issued update to the serving endpoint. */
+  servedModels?: ServedModel[] | undefined;
+  /** The traffic config defining how invocations to the serving endpoint should be routed. */
+  trafficConfig?: TrafficConfig | undefined;
+  /** The config version that the serving endpoint is currently serving. */
+  configVersion?: number | undefined;
+  /** The timestamp when the update to the pending config started. */
+  startTime?: bigint | undefined;
+  /**
+   * Configuration for legacy Inference Tables which automatically log requests and responses to Unity
+   * Catalog.
+   * Deprecated: please use AI Gateway inference tables instead. See
+   * https://docs.databricks.com/aws/en/ai-gateway/inference-tables.
+   */
+  autoCaptureConfig?: AutoCaptureConfig | undefined;
+}
+
+export interface PiiSettings {
+  /** Configuration for input guardrail filters. */
+  behavior?: Behavior | undefined;
+}
+
+export interface PtEndpointCoreConfig {
+  /** The list of served entities under the serving endpoint config. */
+  servedEntities?: PtServedModel[] | undefined;
+  trafficConfig?: TrafficConfig | undefined;
+}
+
+export interface PtServedModel {
+  /** The name of a served entity. It must be unique across an endpoint. A served entity name can consist of alphanumeric characters, dashes, and underscores. If not specified for an external model, this field defaults to external_model.name, with '.' and ':' replaced with '-', and if not specified for other entities, it defaults to entity_name-entity_version. */
+  name?: string | undefined;
+  /** The name of the entity to be served. The entity may be a model in the Databricks Model Registry, a model in the Unity Catalog (UC), or a function of type FEATURE_SPEC in the UC. If it is a UC object, the full name of the object should be given in the form of **catalog_name.schema_name.model_name**. */
+  entityName?: string | undefined;
+  entityVersion?: string | undefined;
+  /** The number of model units to be provisioned. */
+  provisionedModelUnits?: bigint | undefined;
+  /**
+   * Whether burst scaling is enabled. When enabled (default), the endpoint can automatically
+   * scale up beyond provisioned capacity to handle traffic spikes. When disabled, the endpoint
+   * maintains fixed capacity at provisioned_model_units.
+   */
+  burstScalingEnabled?: boolean | undefined;
+}
+
+export interface PutInferenceEndpointAiGatewayRequest {
+  /** The name of the serving endpoint whose AI Gateway is being updated. This field is required. */
+  name?: string | undefined;
+  /**
+   * Configuration to enable usage tracking using system tables.
+   * These tables allow you to monitor operational usage on endpoints and their associated costs.
+   */
+  usageTrackingConfig?: UsageTrackingConfig | undefined;
+  /**
+   * Configuration for payload logging using inference tables.
+   * Use these tables to monitor and audit data being sent to and received from model APIs and to improve model quality.
+   */
+  inferenceTableConfig?: InferenceTableConfig | undefined;
+  /** Configuration for rate limits which can be set to limit endpoint traffic. */
+  rateLimits?: AiGatewayRateLimit[] | undefined;
+  /** Configuration for AI Guardrails to prevent unwanted data and unsafe data in requests and responses. */
+  guardrails?: AiGuardrails | undefined;
+  /**
+   * Configuration for traffic fallback which auto fallbacks to other served entities if the request to a served
+   * entity fails with certain error codes, to increase availability.
+   */
+  fallbackConfig?: FallbackConfig | undefined;
+}
+
+// eslint-disable-next-line @typescript-eslint/naming-convention -- Proto-style nested message name.
+export interface PutInferenceEndpointAiGatewayRequest_Response {
+  /**
+   * Configuration to enable usage tracking using system tables.
+   * These tables allow you to monitor operational usage on endpoints and their associated costs.
+   */
+  usageTrackingConfig?: UsageTrackingConfig | undefined;
+  /**
+   * Configuration for payload logging using inference tables.
+   * Use these tables to monitor and audit data being sent to and received from model APIs and to improve model quality.
+   */
+  inferenceTableConfig?: InferenceTableConfig | undefined;
+  /** Configuration for rate limits which can be set to limit endpoint traffic. */
+  rateLimits?: AiGatewayRateLimit[] | undefined;
+  /** Configuration for AI Guardrails to prevent unwanted data and unsafe data in requests and responses. */
+  guardrails?: AiGuardrails | undefined;
+  /**
+   * Configuration for traffic fallback which auto fallbacks to other served entities if the request to a served
+   * entity fails with certain error codes, to increase availability.
+   */
+  fallbackConfig?: FallbackConfig | undefined;
+}
+
+export interface PutInferenceEndpointConfigRequest {
+  /** The name of the serving endpoint to update. This field is required. */
+  name?: string | undefined;
+  /** The list of served entities under the serving endpoint config. */
+  servedEntities?: ServedModel[] | undefined;
+  /** (Deprecated, use served_entities instead) The list of served models under the serving endpoint config. */
+  servedModels?: ServedModel[] | undefined;
+  /** The traffic configuration associated with the serving endpoint config. */
+  trafficConfig?: TrafficConfig | undefined;
+  /**
+   * Configuration for legacy Inference Tables which automatically log requests and responses to Unity
+   * Catalog.
+   * Deprecated: please use AI Gateway inference tables instead. See
+   * https://docs.databricks.com/aws/en/ai-gateway/inference-tables.
+   */
+  autoCaptureConfig?: AutoCaptureConfig | undefined;
+}
+
+export interface PutInferenceEndpointRateLimitsRequest {
+  /** The name of the serving endpoint whose rate limits are being updated. This field is required. */
+  name?: string | undefined;
+  /** The list of endpoint rate limits. */
+  rateLimits?: RateLimit[] | undefined;
+}
+
+// eslint-disable-next-line @typescript-eslint/naming-convention -- Proto-style nested message name.
+export interface PutInferenceEndpointRateLimitsRequest_Response {
+  /** The list of endpoint rate limits. */
+  rateLimits?: RateLimit[] | undefined;
+}
+
+export interface PutPtEndpointConfigRequest {
+  /** The name of the pt endpoint to update. This field is required. */
+  name?: string | undefined;
+  config?: PtEndpointCoreConfig | undefined;
+}
+
+export interface RateLimit {
+  /** Used to specify how many calls are allowed for a key within the renewal_period. */
+  calls?: bigint | undefined;
+  /** Key field for a serving endpoint rate limit. Currently, only 'user' and 'endpoint' are supported, with 'endpoint' being the default if not specified. */
+  key?: string | undefined;
+  /** Renewal period field for a serving endpoint rate limit. Currently, only 'minute' is supported. */
+  renewalPeriod?: string | undefined;
+}
+
+export interface Route {
+  /** The name of the served model this route configures traffic for. */
+  servedModelName?: string | undefined;
+  /** The percentage of endpoint traffic to send to this route. It must be an integer between 0 and 100 inclusive. */
+  trafficPercentage?: number | undefined;
+  servedEntityName?: string | undefined;
+}
+
+export interface ServedModel {
+  /** The name of a served entity. It must be unique across an endpoint. A served entity name can consist of alphanumeric characters, dashes, and underscores. If not specified for an external model, this field defaults to external_model.name, with '.' and ':' replaced with '-', and if not specified for other entities, it defaults to entity_name-entity_version. */
+  name?: string | undefined;
+  /** The external model to be served. NOTE: Only one of external_model and (entity_name, entity_version, workload_size, workload_type, and scale_to_zero_enabled) can be specified with the latter set being used for custom model serving for a <Databricks> registered model. For an existing endpoint with external_model, it cannot be updated to an endpoint without external_model. If the endpoint is created without external_model, users cannot update it to add external_model later. The task type of all external models within an endpoint must be the same. */
+  externalModel?: ExternalModel | undefined;
+  /** The name of the entity to be served. The entity may be a model in the Databricks Model Registry, a model in the Unity Catalog (UC), or a function of type FEATURE_SPEC in the UC. If it is a UC object, the full name of the object should be given in the form of **catalog_name.schema_name.model_name**. */
+  entityName?: string | undefined;
+  entityVersion?: string | undefined;
+  /** The minimum tokens per second that the endpoint can scale down to. */
+  minProvisionedThroughput?: number | undefined;
+  /** The maximum tokens per second that the endpoint can scale up to. */
+  maxProvisionedThroughput?: number | undefined;
+  /** The minimum provisioned concurrency that the endpoint can scale down to. Do not use if workload_size is specified. */
+  minProvisionedConcurrency?: number | undefined;
+  /** The maximum provisioned concurrency that the endpoint can scale up to. Do not use if workload_size is specified. */
+  maxProvisionedConcurrency?: number | undefined;
+  /** The workload size of the served entity. The workload size corresponds to a range of provisioned concurrency that the compute autoscales between. A single unit of provisioned concurrency can process one request at a time. Valid workload sizes are "Small" (4 - 4 provisioned concurrency), "Medium" (8 - 16 provisioned concurrency), and "Large" (16 - 64 provisioned concurrency). Additional custom workload sizes can also be used when available in the workspace. If scale-to-zero is enabled, the lower bound of the provisioned concurrency for each workload size is 0. Do not use if min_provisioned_concurrency and max_provisioned_concurrency are specified. */
+  workloadSize?: string | undefined;
+  /** The number of model units provisioned. */
+  provisionedModelUnits?: bigint | undefined;
+  /**
+   * Whether burst scaling is enabled. When enabled (default), the endpoint can automatically
+   * scale up beyond provisioned capacity to handle traffic spikes. When disabled, the endpoint
+   * maintains fixed capacity at provisioned_model_units.
+   */
+  burstScalingEnabled?: boolean | undefined;
+  /** Whether the compute resources for the served entity should scale down to zero. */
+  scaleToZeroEnabled?: boolean | undefined;
+  modelName?: string | undefined;
+  modelVersion?: string | undefined;
+  /** An object containing a set of optional, user-specified environment variable key-value pairs used for serving this entity. Note: this is an experimental feature and subject to change. Example entity environment variables that refer to <Databricks> secrets: `{"OPENAI_API_KEY": "{{secrets/my_scope/my_key}}", "DATABRICKS_TOKEN": "{{secrets/my_scope2/my_key2}}"}` */
+  environmentVars?: Record<string, string> | undefined;
+  /** ARN of the instance profile that the served entity uses to access AWS resources. */
+  instanceProfileArn?: string | undefined;
+  foundationModel?: FoundationModel | undefined;
+  state?: ServedModelState | undefined;
+  creator?: string | undefined;
+  creationTimestamp?: bigint | undefined;
+}
+
+// eslint-disable-next-line @typescript-eslint/naming-convention -- Proto-style nested message name.
+export interface ServedModel_EnvironmentVarsEntry {
+  /** Name of the environment variable (e.g., OPENAI_API_KEY). */
+  key?: string | undefined;
+  /**
+   * Value of the environment variable. To set secret environment variable that loads and uses
+   * databricks secret, the value format must be "{{secrets/secret_scope/secret_key}}".
+   * The secret_scope must be replaced with the scope of the secret and secret_key
+   * must be replaced with the key of the secret.
+   */
+  value?: string | undefined;
+}
+
+export interface ServedModelLite {
+  name?: string | undefined;
+  /** Only one of model_name and entity_name should be populated */
+  modelName?: string | undefined;
+  entityName?: string | undefined;
+  /** Only one of model_version and entity_version should be populated */
+  modelVersion?: string | undefined;
+  entityVersion?: string | undefined;
+  externalModel?: ExternalModel | undefined;
+  foundationModel?: FoundationModel | undefined;
+}
+
+export interface ServedModelState {
+  deployment?: ServedModelDeploymentState | undefined;
+  deploymentStateMessage?: string | undefined;
+}
+
+export interface TrafficConfig {
+  /** The list of routes that define traffic to each served entity. */
+  routes?: Route[] | undefined;
+}
+
+export interface UpdateInferenceEndpointNotificationsRequest {
+  /** The name of the serving endpoint whose notifications are being updated. This field is required. */
+  name?: string | undefined;
+  /** The email notification settings to update. Specify email addresses to notify when endpoint state changes occur. */
+  emailNotifications?: EmailNotifications | undefined;
+}
+
+// eslint-disable-next-line @typescript-eslint/naming-convention -- Proto-style nested message name.
+export interface UpdateInferenceEndpointNotificationsRequest_Response {
+  name?: string | undefined;
+  emailNotifications?: EmailNotifications | undefined;
+}
+
+export interface UsageTrackingConfig {
+  /** Whether to enable usage tracking. */
+  enabled?: boolean | undefined;
+}
+
+export const unmarshalAi21LabsConfigSchema: z.ZodType<Ai21LabsConfig> = z
+  .object({
+    ai21labs_api_key: z.string().optional(),
+    ai21labs_api_key_plaintext: z.string().optional(),
+  })
+  .transform(d => ({
+    ai21labsApiKey: d.ai21labs_api_key,
+    ai21labsApiKeyPlaintext: d.ai21labs_api_key_plaintext,
+  }));
+
+export const unmarshalAiGatewayConfigSchema: z.ZodType<AiGatewayConfig> = z
+  .object({
+    usage_tracking_config: z
+      .lazy(() => unmarshalUsageTrackingConfigSchema)
+      .optional(),
+    inference_table_config: z
+      .lazy(() => unmarshalInferenceTableConfigSchema)
+      .optional(),
+    rate_limits: z
+      .array(z.lazy(() => unmarshalAiGatewayRateLimitSchema))
+      .optional(),
+    guardrails: z.lazy(() => unmarshalAiGuardrailsSchema).optional(),
+    fallback_config: z.lazy(() => unmarshalFallbackConfigSchema).optional(),
+  })
+  .transform(d => ({
+    usageTrackingConfig: d.usage_tracking_config,
+    inferenceTableConfig: d.inference_table_config,
+    rateLimits: d.rate_limits,
+    guardrails: d.guardrails,
+    fallbackConfig: d.fallback_config,
+  }));
+
+export const unmarshalAiGatewayRateLimitSchema: z.ZodType<AiGatewayRateLimit> =
+  z
+    .object({
+      calls: z
+        .union([z.number(), z.bigint()])
+        .transform(v => BigInt(v))
+        .optional(),
+      key: z.string().optional(),
+      renewal_period: z.string().optional(),
+      principal: z.string().optional(),
+      tokens: z
+        .union([z.number(), z.bigint()])
+        .transform(v => BigInt(v))
+        .optional(),
+    })
+    .transform(d => ({
+      calls: d.calls,
+      key: d.key,
+      renewalPeriod: d.renewal_period,
+      principal: d.principal,
+      tokens: d.tokens,
+    }));
+
+export const unmarshalAiGuardrailParametersSchema: z.ZodType<AiGuardrailParameters> =
+  z
+    .object({
+      safety: z.boolean().optional(),
+      pii: z.lazy(() => unmarshalPiiSettingsSchema).optional(),
+      valid_topics: z.array(z.string()).optional(),
+      invalid_keywords: z.array(z.string()).optional(),
+    })
+    .transform(d => ({
+      safety: d.safety,
+      pii: d.pii,
+      validTopics: d.valid_topics,
+      invalidKeywords: d.invalid_keywords,
+    }));
+
+export const unmarshalAiGuardrailsSchema: z.ZodType<AiGuardrails> = z
+  .object({
+    input: z.lazy(() => unmarshalAiGuardrailParametersSchema).optional(),
+    output: z.lazy(() => unmarshalAiGuardrailParametersSchema).optional(),
+  })
+  .transform(d => ({
+    input: d.input,
+    output: d.output,
+  }));
+
+export const unmarshalAmazonBedrockConfigSchema: z.ZodType<AmazonBedrockConfig> =
+  z
+    .object({
+      aws_region: z.string().optional(),
+      aws_access_key_id: z.string().optional(),
+      aws_secret_access_key: z.string().optional(),
+      bedrock_provider: z.string().optional(),
+      aws_access_key_id_plaintext: z.string().optional(),
+      aws_secret_access_key_plaintext: z.string().optional(),
+      instance_profile_arn: z.string().optional(),
+    })
+    .transform(d => ({
+      awsRegion: d.aws_region,
+      awsAccessKeyId: d.aws_access_key_id,
+      awsSecretAccessKey: d.aws_secret_access_key,
+      bedrockProvider: d.bedrock_provider,
+      awsAccessKeyIdPlaintext: d.aws_access_key_id_plaintext,
+      awsSecretAccessKeyPlaintext: d.aws_secret_access_key_plaintext,
+      instanceProfileArn: d.instance_profile_arn,
+    }));
+
+export const unmarshalAnthropicConfigSchema: z.ZodType<AnthropicConfig> = z
+  .object({
+    anthropic_api_key: z.string().optional(),
+    anthropic_api_key_plaintext: z.string().optional(),
+  })
+  .transform(d => ({
+    anthropicApiKey: d.anthropic_api_key,
+    anthropicApiKeyPlaintext: d.anthropic_api_key_plaintext,
+  }));
+
+export const unmarshalApiKeyAuthSchema: z.ZodType<ApiKeyAuth> = z
+  .object({
+    key: z.string().optional(),
+    value: z.string().optional(),
+    value_plaintext: z.string().optional(),
+  })
+  .transform(d => ({
+    key: d.key,
+    value: d.value,
+    valuePlaintext: d.value_plaintext,
+  }));
+
+export const unmarshalAutoCaptureConfigSchema: z.ZodType<AutoCaptureConfig> = z
+  .object({
+    catalog_name: z.string().optional(),
+    schema_name: z.string().optional(),
+    table_name_prefix: z.string().optional(),
+    state: z.lazy(() => unmarshalAutoCaptureStateSchema).optional(),
+    enabled: z.boolean().optional(),
+  })
+  .transform(d => ({
+    catalogName: d.catalog_name,
+    schemaName: d.schema_name,
+    tableNamePrefix: d.table_name_prefix,
+    state: d.state,
+    enabled: d.enabled,
+  }));
+
+export const unmarshalAutoCaptureStateSchema: z.ZodType<AutoCaptureState> = z
+  .object({
+    payload_table: z.lazy(() => unmarshalPayloadTableSchema).optional(),
+  })
+  .transform(d => ({
+    payloadTable: d.payload_table,
+  }));
+
+export const unmarshalBearerTokenAuthSchema: z.ZodType<BearerTokenAuth> = z
+  .object({
+    token: z.string().optional(),
+    token_plaintext: z.string().optional(),
+  })
+  .transform(d => ({
+    token: d.token,
+    tokenPlaintext: d.token_plaintext,
+  }));
+
+export const unmarshalCohereConfigSchema: z.ZodType<CohereConfig> = z
+  .object({
+    cohere_api_key: z.string().optional(),
+    cohere_api_key_plaintext: z.string().optional(),
+    cohere_api_base: z.string().optional(),
+  })
+  .transform(d => ({
+    cohereApiKey: d.cohere_api_key,
+    cohereApiKeyPlaintext: d.cohere_api_key_plaintext,
+    cohereApiBase: d.cohere_api_base,
+  }));
+
+export const unmarshalCustomProviderConfigSchema: z.ZodType<CustomProviderConfig> =
+  z
+    .object({
+      custom_provider_url: z.string().optional(),
+      bearer_token_auth: z
+        .lazy(() => unmarshalBearerTokenAuthSchema)
+        .optional(),
+      api_key_auth: z.lazy(() => unmarshalApiKeyAuthSchema).optional(),
+    })
+    .transform(d => ({
+      customProviderUrl: d.custom_provider_url,
+      bearerTokenAuth: d.bearer_token_auth,
+      apiKeyAuth: d.api_key_auth,
+    }));
+
+export const unmarshalDataPlaneInfoSchema: z.ZodType<DataPlaneInfo> = z
+  .object({
+    endpoint_url: z.string().optional(),
+    authorization_details: z.string().optional(),
+  })
+  .transform(d => ({
+    endpointUrl: d.endpoint_url,
+    authorizationDetails: d.authorization_details,
+  }));
+
+export const unmarshalDatabricksModelServingConfigSchema: z.ZodType<DatabricksModelServingConfig> =
+  z
+    .object({
+      databricks_api_token: z.string().optional(),
+      databricks_workspace_url: z.string().optional(),
+      databricks_api_token_plaintext: z.string().optional(),
+    })
+    .transform(d => ({
+      databricksApiToken: d.databricks_api_token,
+      databricksWorkspaceUrl: d.databricks_workspace_url,
+      databricksApiTokenPlaintext: d.databricks_api_token_plaintext,
+    }));
+
+// eslint-disable-next-line @typescript-eslint/naming-convention -- Proto-style nested message name.
+export const unmarshalDeleteInferenceEndpointRequest_ResponseSchema: z.ZodType<DeleteInferenceEndpointRequest_Response> =
+  z.object({});
+
+export const unmarshalEmailNotificationsSchema: z.ZodType<EmailNotifications> =
+  z
+    .object({
+      on_update_success: z.array(z.string()).optional(),
+      on_update_failure: z.array(z.string()).optional(),
+    })
+    .transform(d => ({
+      onUpdateSuccess: d.on_update_success,
+      onUpdateFailure: d.on_update_failure,
+    }));
+
+export const unmarshalEndpointCoreConfigOutputSchema: z.ZodType<EndpointCoreConfigOutput> =
+  z
+    .object({
+      config_version: z
+        .union([z.number(), z.bigint()])
+        .transform(v => BigInt(v))
+        .optional(),
+      served_entities: z
+        .array(z.lazy(() => unmarshalServedModelSchema))
+        .optional(),
+      served_models: z
+        .array(z.lazy(() => unmarshalServedModelSchema))
+        .optional(),
+      traffic_config: z.lazy(() => unmarshalTrafficConfigSchema).optional(),
+      auto_capture_config: z
+        .lazy(() => unmarshalAutoCaptureConfigSchema)
+        .optional(),
+    })
+    .transform(d => ({
+      configVersion: d.config_version,
+      servedEntities: d.served_entities,
+      servedModels: d.served_models,
+      trafficConfig: d.traffic_config,
+      autoCaptureConfig: d.auto_capture_config,
+    }));
+
+export const unmarshalEndpointCoreConfigSummarySchema: z.ZodType<EndpointCoreConfigSummary> =
+  z
+    .object({
+      served_entities: z
+        .array(z.lazy(() => unmarshalServedModelLiteSchema))
+        .optional(),
+      served_models: z
+        .array(z.lazy(() => unmarshalServedModelLiteSchema))
+        .optional(),
+    })
+    .transform(d => ({
+      servedEntities: d.served_entities,
+      servedModels: d.served_models,
+    }));
+
+export const unmarshalEndpointTagSchema: z.ZodType<EndpointTag> = z
+  .object({
+    key: z.string().optional(),
+    value: z.string().optional(),
+  })
+  .transform(d => ({
+    key: d.key,
+    value: d.value,
+  }));
+
+export const unmarshalExternalModelSchema: z.ZodType<ExternalModel> = z
+  .object({
+    provider: z.string().optional(),
+    name: z.string().optional(),
+    task: z.string().optional(),
+    ai21labs_config: z.lazy(() => unmarshalAi21LabsConfigSchema).optional(),
+    anthropic_config: z.lazy(() => unmarshalAnthropicConfigSchema).optional(),
+    amazon_bedrock_config: z
+      .lazy(() => unmarshalAmazonBedrockConfigSchema)
+      .optional(),
+    cohere_config: z.lazy(() => unmarshalCohereConfigSchema).optional(),
+    google_cloud_vertex_ai_config: z
+      .lazy(() => unmarshalGoogleCloudVertexAiConfigSchema)
+      .optional(),
+    databricks_model_serving_config: z
+      .lazy(() => unmarshalDatabricksModelServingConfigSchema)
+      .optional(),
+    openai_config: z.lazy(() => unmarshalOpenAiConfigSchema).optional(),
+    palm_config: z.lazy(() => unmarshalPaLmConfigSchema).optional(),
+    custom_provider_config: z
+      .lazy(() => unmarshalCustomProviderConfigSchema)
+      .optional(),
+  })
+  .transform(d => ({
+    provider: d.provider,
+    name: d.name,
+    task: d.task,
+    config:
+      d.ai21labs_config !== undefined
+        ? {$case: 'ai21labsConfig' as const, ai21labsConfig: d.ai21labs_config}
+        : d.anthropic_config !== undefined
+          ? {
+              $case: 'anthropicConfig' as const,
+              anthropicConfig: d.anthropic_config,
+            }
+          : d.amazon_bedrock_config !== undefined
+            ? {
+                $case: 'amazonBedrockConfig' as const,
+                amazonBedrockConfig: d.amazon_bedrock_config,
+              }
+            : d.cohere_config !== undefined
+              ? {$case: 'cohereConfig' as const, cohereConfig: d.cohere_config}
+              : d.google_cloud_vertex_ai_config !== undefined
+                ? {
+                    $case: 'googleCloudVertexAiConfig' as const,
+                    googleCloudVertexAiConfig: d.google_cloud_vertex_ai_config,
+                  }
+                : d.databricks_model_serving_config !== undefined
+                  ? {
+                      $case: 'databricksModelServingConfig' as const,
+                      databricksModelServingConfig:
+                        d.databricks_model_serving_config,
+                    }
+                  : d.openai_config !== undefined
+                    ? {
+                        $case: 'openaiConfig' as const,
+                        openaiConfig: d.openai_config,
+                      }
+                    : d.palm_config !== undefined
+                      ? {
+                          $case: 'palmConfig' as const,
+                          palmConfig: d.palm_config,
+                        }
+                      : d.custom_provider_config !== undefined
+                        ? {
+                            $case: 'customProviderConfig' as const,
+                            customProviderConfig: d.custom_provider_config,
+                          }
+                        : undefined,
+  }));
+
+export const unmarshalFallbackConfigSchema: z.ZodType<FallbackConfig> = z
+  .object({
+    enabled: z.boolean().optional(),
+  })
+  .transform(d => ({
+    enabled: d.enabled,
+  }));
+
+export const unmarshalFoundationModelSchema: z.ZodType<FoundationModel> = z
+  .object({
+    name: z.string().optional(),
+    display_name: z.string().optional(),
+    docs: z.string().optional(),
+    description: z.string().optional(),
+  })
+  .transform(d => ({
+    name: d.name,
+    displayName: d.display_name,
+    docs: d.docs,
+    description: d.description,
+  }));
+
+// eslint-disable-next-line @typescript-eslint/naming-convention -- Proto-style nested message name.
+export const unmarshalGetServedModelBuildLogsRequest_ResponseSchema: z.ZodType<GetServedModelBuildLogsRequest_Response> =
+  z
+    .object({
+      logs: z.string().optional(),
+    })
+    .transform(d => ({
+      logs: d.logs,
+    }));
+
+// eslint-disable-next-line @typescript-eslint/naming-convention -- Proto-style nested message name.
+export const unmarshalGetServedModelLogsRequest_ResponseSchema: z.ZodType<GetServedModelLogsRequest_Response> =
+  z
+    .object({
+      logs: z.string().optional(),
+    })
+    .transform(d => ({
+      logs: d.logs,
+    }));
+
+export const unmarshalGoogleCloudVertexAiConfigSchema: z.ZodType<GoogleCloudVertexAiConfig> =
+  z
+    .object({
+      private_key: z.string().optional(),
+      project_id: z.string().optional(),
+      region: z.string().optional(),
+      private_key_plaintext: z.string().optional(),
+    })
+    .transform(d => ({
+      privateKey: d.private_key,
+      projectId: d.project_id,
+      region: d.region,
+      privateKeyPlaintext: d.private_key_plaintext,
+    }));
+
+export const unmarshalInferenceEndpointSchema: z.ZodType<InferenceEndpoint> = z
+  .object({
+    name: z.string().optional(),
+    creator: z.string().optional(),
+    creation_timestamp: z
+      .union([z.number(), z.bigint()])
+      .transform(v => BigInt(v))
+      .optional(),
+    last_updated_timestamp: z
+      .union([z.number(), z.bigint()])
+      .transform(v => BigInt(v))
+      .optional(),
+    state: z.lazy(() => unmarshalInferenceEndpointStateSchema).optional(),
+    config: z.lazy(() => unmarshalEndpointCoreConfigSummarySchema).optional(),
+    tags: z.array(z.lazy(() => unmarshalEndpointTagSchema)).optional(),
+    id: z.string().optional(),
+    task: z.string().optional(),
+    ai_gateway: z.lazy(() => unmarshalAiGatewayConfigSchema).optional(),
+    budget_policy_id: z.string().optional(),
+    description: z.string().optional(),
+    usage_policy_id: z.string().optional(),
+  })
+  .transform(d => ({
+    name: d.name,
+    creator: d.creator,
+    creationTimestamp: d.creation_timestamp,
+    lastUpdatedTimestamp: d.last_updated_timestamp,
+    state: d.state,
+    config: d.config,
+    tags: d.tags,
+    id: d.id,
+    task: d.task,
+    aiGateway: d.ai_gateway,
+    budgetPolicyId: d.budget_policy_id,
+    description: d.description,
+    usagePolicyId: d.usage_policy_id,
+  }));
+
+export const unmarshalInferenceEndpointDetailedSchema: z.ZodType<InferenceEndpointDetailed> =
+  z
+    .object({
+      name: z.string().optional(),
+      creator: z.string().optional(),
+      creation_timestamp: z
+        .union([z.number(), z.bigint()])
+        .transform(v => BigInt(v))
+        .optional(),
+      last_updated_timestamp: z
+        .union([z.number(), z.bigint()])
+        .transform(v => BigInt(v))
+        .optional(),
+      state: z.lazy(() => unmarshalInferenceEndpointStateSchema).optional(),
+      config: z.lazy(() => unmarshalEndpointCoreConfigOutputSchema).optional(),
+      pending_config: z.lazy(() => unmarshalPendingConfigSchema).optional(),
+      id: z.string().optional(),
+      permission_level: z
+        .enum(ServingEndpointDetailedPermissionLevel)
+        .optional(),
+      tags: z.array(z.lazy(() => unmarshalEndpointTagSchema)).optional(),
+      task: z.string().optional(),
+      route_optimized: z.boolean().optional(),
+      endpoint_url: z.string().optional(),
+      data_plane_info: z
+        .lazy(() => unmarshalModelDataPlaneInfoSchema)
+        .optional(),
+      ai_gateway: z.lazy(() => unmarshalAiGatewayConfigSchema).optional(),
+      budget_policy_id: z.string().optional(),
+      email_notifications: z
+        .lazy(() => unmarshalEmailNotificationsSchema)
+        .optional(),
+      description: z.string().optional(),
+    })
+    .transform(d => ({
+      name: d.name,
+      creator: d.creator,
+      creationTimestamp: d.creation_timestamp,
+      lastUpdatedTimestamp: d.last_updated_timestamp,
+      state: d.state,
+      config: d.config,
+      pendingConfig: d.pending_config,
+      id: d.id,
+      permissionLevel: d.permission_level,
+      tags: d.tags,
+      task: d.task,
+      routeOptimized: d.route_optimized,
+      endpointUrl: d.endpoint_url,
+      dataPlaneInfo: d.data_plane_info,
+      aiGateway: d.ai_gateway,
+      budgetPolicyId: d.budget_policy_id,
+      emailNotifications: d.email_notifications,
+      description: d.description,
+    }));
+
+export const unmarshalInferenceEndpointStateSchema: z.ZodType<InferenceEndpointState> =
+  z
+    .object({
+      ready: z.enum(InferenceEndpointState_ReadyState).optional(),
+      config_update: z
+        .enum(InferenceEndpointState_ConfigUpdateState)
+        .optional(),
+    })
+    .transform(d => ({
+      ready: d.ready,
+      configUpdate: d.config_update,
+    }));
+
+export const unmarshalInferenceTableConfigSchema: z.ZodType<InferenceTableConfig> =
+  z
+    .object({
+      catalog_name: z.string().optional(),
+      schema_name: z.string().optional(),
+      table_name_prefix: z.string().optional(),
+      enabled: z.boolean().optional(),
+    })
+    .transform(d => ({
+      catalogName: d.catalog_name,
+      schemaName: d.schema_name,
+      tableNamePrefix: d.table_name_prefix,
+      enabled: d.enabled,
+    }));
+
+// eslint-disable-next-line @typescript-eslint/naming-convention -- Proto-style nested message name.
+export const unmarshalListInferenceEndpointsRequest_ResponseSchema: z.ZodType<ListInferenceEndpointsRequest_Response> =
+  z
+    .object({
+      endpoints: z
+        .array(z.lazy(() => unmarshalInferenceEndpointSchema))
+        .optional(),
+    })
+    .transform(d => ({
+      endpoints: d.endpoints,
+    }));
+
+export const unmarshalModelDataPlaneInfoSchema: z.ZodType<ModelDataPlaneInfo> =
+  z
+    .object({
+      query_info: z.lazy(() => unmarshalDataPlaneInfoSchema).optional(),
+    })
+    .transform(d => ({
+      queryInfo: d.query_info,
+    }));
+
+export const unmarshalOpenAiConfigSchema: z.ZodType<OpenAiConfig> = z
+  .object({
+    openai_api_key: z.string().optional(),
+    openai_api_type: z.string().optional(),
+    openai_api_base: z.string().optional(),
+    openai_api_version: z.string().optional(),
+    openai_deployment_name: z.string().optional(),
+    openai_organization: z.string().optional(),
+    microsoft_entra_tenant_id: z.string().optional(),
+    microsoft_entra_client_id: z.string().optional(),
+    microsoft_entra_client_secret: z.string().optional(),
+    openai_api_key_plaintext: z.string().optional(),
+    microsoft_entra_client_secret_plaintext: z.string().optional(),
+  })
+  .transform(d => ({
+    openaiApiKey: d.openai_api_key,
+    openaiApiType: d.openai_api_type,
+    openaiApiBase: d.openai_api_base,
+    openaiApiVersion: d.openai_api_version,
+    openaiDeploymentName: d.openai_deployment_name,
+    openaiOrganization: d.openai_organization,
+    microsoftEntraTenantId: d.microsoft_entra_tenant_id,
+    microsoftEntraClientId: d.microsoft_entra_client_id,
+    microsoftEntraClientSecret: d.microsoft_entra_client_secret,
+    openaiApiKeyPlaintext: d.openai_api_key_plaintext,
+    microsoftEntraClientSecretPlaintext:
+      d.microsoft_entra_client_secret_plaintext,
+  }));
+
+export const unmarshalPaLmConfigSchema: z.ZodType<PaLmConfig> = z
+  .object({
+    palm_api_key: z.string().optional(),
+    palm_api_key_plaintext: z.string().optional(),
+  })
+  .transform(d => ({
+    palmApiKey: d.palm_api_key,
+    palmApiKeyPlaintext: d.palm_api_key_plaintext,
+  }));
+
+// eslint-disable-next-line @typescript-eslint/naming-convention -- Proto-style nested message name.
+export const unmarshalPatchInferenceEndpointTagsRequest_ResponseSchema: z.ZodType<PatchInferenceEndpointTagsRequest_Response> =
+  z
+    .object({
+      tags: z.array(z.lazy(() => unmarshalEndpointTagSchema)).optional(),
+    })
+    .transform(d => ({
+      tags: d.tags,
+    }));
+
+export const unmarshalPayloadTableSchema: z.ZodType<PayloadTable> = z
+  .object({
+    name: z.string().optional(),
+    status: z.string().optional(),
+    status_message: z.string().optional(),
+  })
+  .transform(d => ({
+    name: d.name,
+    status: d.status,
+    statusMessage: d.status_message,
+  }));
+
+export const unmarshalPendingConfigSchema: z.ZodType<PendingConfig> = z
+  .object({
+    served_entities: z
+      .array(z.lazy(() => unmarshalServedModelSchema))
+      .optional(),
+    served_models: z.array(z.lazy(() => unmarshalServedModelSchema)).optional(),
+    traffic_config: z.lazy(() => unmarshalTrafficConfigSchema).optional(),
+    config_version: z.number().optional(),
+    start_time: z
+      .union([z.number(), z.bigint()])
+      .transform(v => BigInt(v))
+      .optional(),
+    auto_capture_config: z
+      .lazy(() => unmarshalAutoCaptureConfigSchema)
+      .optional(),
+  })
+  .transform(d => ({
+    servedEntities: d.served_entities,
+    servedModels: d.served_models,
+    trafficConfig: d.traffic_config,
+    configVersion: d.config_version,
+    startTime: d.start_time,
+    autoCaptureConfig: d.auto_capture_config,
+  }));
+
+export const unmarshalPiiSettingsSchema: z.ZodType<PiiSettings> = z
+  .object({
+    behavior: z.enum(Behavior).optional(),
+  })
+  .transform(d => ({
+    behavior: d.behavior,
+  }));
+
+// eslint-disable-next-line @typescript-eslint/naming-convention -- Proto-style nested message name.
+export const unmarshalPutInferenceEndpointAiGatewayRequest_ResponseSchema: z.ZodType<PutInferenceEndpointAiGatewayRequest_Response> =
+  z
+    .object({
+      usage_tracking_config: z
+        .lazy(() => unmarshalUsageTrackingConfigSchema)
+        .optional(),
+      inference_table_config: z
+        .lazy(() => unmarshalInferenceTableConfigSchema)
+        .optional(),
+      rate_limits: z
+        .array(z.lazy(() => unmarshalAiGatewayRateLimitSchema))
+        .optional(),
+      guardrails: z.lazy(() => unmarshalAiGuardrailsSchema).optional(),
+      fallback_config: z.lazy(() => unmarshalFallbackConfigSchema).optional(),
+    })
+    .transform(d => ({
+      usageTrackingConfig: d.usage_tracking_config,
+      inferenceTableConfig: d.inference_table_config,
+      rateLimits: d.rate_limits,
+      guardrails: d.guardrails,
+      fallbackConfig: d.fallback_config,
+    }));
+
+// eslint-disable-next-line @typescript-eslint/naming-convention -- Proto-style nested message name.
+export const unmarshalPutInferenceEndpointRateLimitsRequest_ResponseSchema: z.ZodType<PutInferenceEndpointRateLimitsRequest_Response> =
+  z
+    .object({
+      rate_limits: z.array(z.lazy(() => unmarshalRateLimitSchema)).optional(),
+    })
+    .transform(d => ({
+      rateLimits: d.rate_limits,
+    }));
+
+export const unmarshalRateLimitSchema: z.ZodType<RateLimit> = z
+  .object({
+    calls: z
+      .union([z.number(), z.bigint()])
+      .transform(v => BigInt(v))
+      .optional(),
+    key: z.string().optional(),
+    renewal_period: z.string().optional(),
+  })
+  .transform(d => ({
+    calls: d.calls,
+    key: d.key,
+    renewalPeriod: d.renewal_period,
+  }));
+
+export const unmarshalRouteSchema: z.ZodType<Route> = z
+  .object({
+    served_model_name: z.string().optional(),
+    traffic_percentage: z.number().optional(),
+    served_entity_name: z.string().optional(),
+  })
+  .transform(d => ({
+    servedModelName: d.served_model_name,
+    trafficPercentage: d.traffic_percentage,
+    servedEntityName: d.served_entity_name,
+  }));
+
+export const unmarshalServedModelSchema: z.ZodType<ServedModel> = z
+  .object({
+    name: z.string().optional(),
+    external_model: z.lazy(() => unmarshalExternalModelSchema).optional(),
+    entity_name: z.string().optional(),
+    entity_version: z.string().optional(),
+    min_provisioned_throughput: z.number().optional(),
+    max_provisioned_throughput: z.number().optional(),
+    min_provisioned_concurrency: z.number().optional(),
+    max_provisioned_concurrency: z.number().optional(),
+    workload_size: z.string().optional(),
+    provisioned_model_units: z
+      .union([z.number(), z.bigint()])
+      .transform(v => BigInt(v))
+      .optional(),
+    burst_scaling_enabled: z.boolean().optional(),
+    scale_to_zero_enabled: z.boolean().optional(),
+    model_name: z.string().optional(),
+    model_version: z.string().optional(),
+    environment_vars: z.record(z.string(), z.string()).optional(),
+    instance_profile_arn: z.string().optional(),
+    foundation_model: z.lazy(() => unmarshalFoundationModelSchema).optional(),
+    state: z.lazy(() => unmarshalServedModelStateSchema).optional(),
+    creator: z.string().optional(),
+    creation_timestamp: z
+      .union([z.number(), z.bigint()])
+      .transform(v => BigInt(v))
+      .optional(),
+  })
+  .transform(d => ({
+    name: d.name,
+    externalModel: d.external_model,
+    entityName: d.entity_name,
+    entityVersion: d.entity_version,
+    minProvisionedThroughput: d.min_provisioned_throughput,
+    maxProvisionedThroughput: d.max_provisioned_throughput,
+    minProvisionedConcurrency: d.min_provisioned_concurrency,
+    maxProvisionedConcurrency: d.max_provisioned_concurrency,
+    workloadSize: d.workload_size,
+    provisionedModelUnits: d.provisioned_model_units,
+    burstScalingEnabled: d.burst_scaling_enabled,
+    scaleToZeroEnabled: d.scale_to_zero_enabled,
+    modelName: d.model_name,
+    modelVersion: d.model_version,
+    environmentVars: d.environment_vars,
+    instanceProfileArn: d.instance_profile_arn,
+    foundationModel: d.foundation_model,
+    state: d.state,
+    creator: d.creator,
+    creationTimestamp: d.creation_timestamp,
+  }));
+
+export const unmarshalServedModelLiteSchema: z.ZodType<ServedModelLite> = z
+  .object({
+    name: z.string().optional(),
+    model_name: z.string().optional(),
+    entity_name: z.string().optional(),
+    model_version: z.string().optional(),
+    entity_version: z.string().optional(),
+    external_model: z.lazy(() => unmarshalExternalModelSchema).optional(),
+    foundation_model: z.lazy(() => unmarshalFoundationModelSchema).optional(),
+  })
+  .transform(d => ({
+    name: d.name,
+    modelName: d.model_name,
+    entityName: d.entity_name,
+    modelVersion: d.model_version,
+    entityVersion: d.entity_version,
+    externalModel: d.external_model,
+    foundationModel: d.foundation_model,
+  }));
+
+export const unmarshalServedModelStateSchema: z.ZodType<ServedModelState> = z
+  .object({
+    deployment: z.enum(ServedModelDeploymentState).optional(),
+    deployment_state_message: z.string().optional(),
+  })
+  .transform(d => ({
+    deployment: d.deployment,
+    deploymentStateMessage: d.deployment_state_message,
+  }));
+
+export const unmarshalTrafficConfigSchema: z.ZodType<TrafficConfig> = z
+  .object({
+    routes: z.array(z.lazy(() => unmarshalRouteSchema)).optional(),
+  })
+  .transform(d => ({
+    routes: d.routes,
+  }));
+
+// eslint-disable-next-line @typescript-eslint/naming-convention -- Proto-style nested message name.
+export const unmarshalUpdateInferenceEndpointNotificationsRequest_ResponseSchema: z.ZodType<UpdateInferenceEndpointNotificationsRequest_Response> =
+  z
+    .object({
+      name: z.string().optional(),
+      email_notifications: z
+        .lazy(() => unmarshalEmailNotificationsSchema)
+        .optional(),
+    })
+    .transform(d => ({
+      name: d.name,
+      emailNotifications: d.email_notifications,
+    }));
+
+export const unmarshalUsageTrackingConfigSchema: z.ZodType<UsageTrackingConfig> =
+  z
+    .object({
+      enabled: z.boolean().optional(),
+    })
+    .transform(d => ({
+      enabled: d.enabled,
+    }));
+
+export const marshalAi21LabsConfigSchema: z.ZodType = z
+  .object({
+    ai21labsApiKey: z.string().optional(),
+    ai21labsApiKeyPlaintext: z.string().optional(),
+  })
+  .transform(d => ({
+    ai21labs_api_key: d.ai21labsApiKey,
+    ai21labs_api_key_plaintext: d.ai21labsApiKeyPlaintext,
+  }));
+
+export const marshalAiGatewayConfigSchema: z.ZodType = z
+  .object({
+    usageTrackingConfig: z
+      .lazy(() => marshalUsageTrackingConfigSchema)
+      .optional(),
+    inferenceTableConfig: z
+      .lazy(() => marshalInferenceTableConfigSchema)
+      .optional(),
+    rateLimits: z
+      .array(z.lazy(() => marshalAiGatewayRateLimitSchema))
+      .optional(),
+    guardrails: z.lazy(() => marshalAiGuardrailsSchema).optional(),
+    fallbackConfig: z.lazy(() => marshalFallbackConfigSchema).optional(),
+  })
+  .transform(d => ({
+    usage_tracking_config: d.usageTrackingConfig,
+    inference_table_config: d.inferenceTableConfig,
+    rate_limits: d.rateLimits,
+    guardrails: d.guardrails,
+    fallback_config: d.fallbackConfig,
+  }));
+
+export const marshalAiGatewayRateLimitSchema: z.ZodType = z
+  .object({
+    calls: z.bigint().optional(),
+    key: z.string().optional(),
+    renewalPeriod: z.string().optional(),
+    principal: z.string().optional(),
+    tokens: z.bigint().optional(),
+  })
+  .transform(d => ({
+    calls: d.calls,
+    key: d.key,
+    renewal_period: d.renewalPeriod,
+    principal: d.principal,
+    tokens: d.tokens,
+  }));
+
+export const marshalAiGuardrailParametersSchema: z.ZodType = z
+  .object({
+    safety: z.boolean().optional(),
+    pii: z.lazy(() => marshalPiiSettingsSchema).optional(),
+    validTopics: z.array(z.string()).optional(),
+    invalidKeywords: z.array(z.string()).optional(),
+  })
+  .transform(d => ({
+    safety: d.safety,
+    pii: d.pii,
+    valid_topics: d.validTopics,
+    invalid_keywords: d.invalidKeywords,
+  }));
+
+export const marshalAiGuardrailsSchema: z.ZodType = z
+  .object({
+    input: z.lazy(() => marshalAiGuardrailParametersSchema).optional(),
+    output: z.lazy(() => marshalAiGuardrailParametersSchema).optional(),
+  })
+  .transform(d => ({
+    input: d.input,
+    output: d.output,
+  }));
+
+export const marshalAmazonBedrockConfigSchema: z.ZodType = z
+  .object({
+    awsRegion: z.string().optional(),
+    awsAccessKeyId: z.string().optional(),
+    awsSecretAccessKey: z.string().optional(),
+    bedrockProvider: z.string().optional(),
+    awsAccessKeyIdPlaintext: z.string().optional(),
+    awsSecretAccessKeyPlaintext: z.string().optional(),
+    instanceProfileArn: z.string().optional(),
+  })
+  .transform(d => ({
+    aws_region: d.awsRegion,
+    aws_access_key_id: d.awsAccessKeyId,
+    aws_secret_access_key: d.awsSecretAccessKey,
+    bedrock_provider: d.bedrockProvider,
+    aws_access_key_id_plaintext: d.awsAccessKeyIdPlaintext,
+    aws_secret_access_key_plaintext: d.awsSecretAccessKeyPlaintext,
+    instance_profile_arn: d.instanceProfileArn,
+  }));
+
+export const marshalAnthropicConfigSchema: z.ZodType = z
+  .object({
+    anthropicApiKey: z.string().optional(),
+    anthropicApiKeyPlaintext: z.string().optional(),
+  })
+  .transform(d => ({
+    anthropic_api_key: d.anthropicApiKey,
+    anthropic_api_key_plaintext: d.anthropicApiKeyPlaintext,
+  }));
+
+export const marshalApiKeyAuthSchema: z.ZodType = z
+  .object({
+    key: z.string().optional(),
+    value: z.string().optional(),
+    valuePlaintext: z.string().optional(),
+  })
+  .transform(d => ({
+    key: d.key,
+    value: d.value,
+    value_plaintext: d.valuePlaintext,
+  }));
+
+export const marshalAutoCaptureConfigSchema: z.ZodType = z
+  .object({
+    catalogName: z.string().optional(),
+    schemaName: z.string().optional(),
+    tableNamePrefix: z.string().optional(),
+    state: z.lazy(() => marshalAutoCaptureStateSchema).optional(),
+    enabled: z.boolean().optional(),
+  })
+  .transform(d => ({
+    catalog_name: d.catalogName,
+    schema_name: d.schemaName,
+    table_name_prefix: d.tableNamePrefix,
+    state: d.state,
+    enabled: d.enabled,
+  }));
+
+export const marshalAutoCaptureStateSchema: z.ZodType = z
+  .object({
+    payloadTable: z.lazy(() => marshalPayloadTableSchema).optional(),
+  })
+  .transform(d => ({
+    payload_table: d.payloadTable,
+  }));
+
+export const marshalBearerTokenAuthSchema: z.ZodType = z
+  .object({
+    token: z.string().optional(),
+    tokenPlaintext: z.string().optional(),
+  })
+  .transform(d => ({
+    token: d.token,
+    token_plaintext: d.tokenPlaintext,
+  }));
+
+export const marshalCohereConfigSchema: z.ZodType = z
+  .object({
+    cohereApiKey: z.string().optional(),
+    cohereApiKeyPlaintext: z.string().optional(),
+    cohereApiBase: z.string().optional(),
+  })
+  .transform(d => ({
+    cohere_api_key: d.cohereApiKey,
+    cohere_api_key_plaintext: d.cohereApiKeyPlaintext,
+    cohere_api_base: d.cohereApiBase,
+  }));
+
+export const marshalCreateInferenceEndpointRequestSchema: z.ZodType = z
+  .object({
+    name: z.string().optional(),
+    config: z.lazy(() => marshalEndpointCoreConfigSchema).optional(),
+    tags: z.array(z.lazy(() => marshalEndpointTagSchema)).optional(),
+    routeOptimized: z.boolean().optional(),
+    rateLimits: z.array(z.lazy(() => marshalRateLimitSchema)).optional(),
+    aiGateway: z.lazy(() => marshalAiGatewayConfigSchema).optional(),
+    budgetPolicyId: z.string().optional(),
+    emailNotifications: z
+      .lazy(() => marshalEmailNotificationsSchema)
+      .optional(),
+    description: z.string().optional(),
+  })
+  .transform(d => ({
+    name: d.name,
+    config: d.config,
+    tags: d.tags,
+    route_optimized: d.routeOptimized,
+    rate_limits: d.rateLimits,
+    ai_gateway: d.aiGateway,
+    budget_policy_id: d.budgetPolicyId,
+    email_notifications: d.emailNotifications,
+    description: d.description,
+  }));
+
+export const marshalCreatePtEndpointRequestSchema: z.ZodType = z
+  .object({
+    name: z.string().optional(),
+    config: z.lazy(() => marshalPtEndpointCoreConfigSchema).optional(),
+    tags: z.array(z.lazy(() => marshalEndpointTagSchema)).optional(),
+    aiGateway: z.lazy(() => marshalAiGatewayConfigSchema).optional(),
+    budgetPolicyId: z.string().optional(),
+    emailNotifications: z
+      .lazy(() => marshalEmailNotificationsSchema)
+      .optional(),
+  })
+  .transform(d => ({
+    name: d.name,
+    config: d.config,
+    tags: d.tags,
+    ai_gateway: d.aiGateway,
+    budget_policy_id: d.budgetPolicyId,
+    email_notifications: d.emailNotifications,
+  }));
+
+export const marshalCustomProviderConfigSchema: z.ZodType = z
+  .object({
+    customProviderUrl: z.string().optional(),
+    bearerTokenAuth: z.lazy(() => marshalBearerTokenAuthSchema).optional(),
+    apiKeyAuth: z.lazy(() => marshalApiKeyAuthSchema).optional(),
+  })
+  .transform(d => ({
+    custom_provider_url: d.customProviderUrl,
+    bearer_token_auth: d.bearerTokenAuth,
+    api_key_auth: d.apiKeyAuth,
+  }));
+
+export const marshalDatabricksModelServingConfigSchema: z.ZodType = z
+  .object({
+    databricksApiToken: z.string().optional(),
+    databricksWorkspaceUrl: z.string().optional(),
+    databricksApiTokenPlaintext: z.string().optional(),
+  })
+  .transform(d => ({
+    databricks_api_token: d.databricksApiToken,
+    databricks_workspace_url: d.databricksWorkspaceUrl,
+    databricks_api_token_plaintext: d.databricksApiTokenPlaintext,
+  }));
+
+export const marshalEmailNotificationsSchema: z.ZodType = z
+  .object({
+    onUpdateSuccess: z.array(z.string()).optional(),
+    onUpdateFailure: z.array(z.string()).optional(),
+  })
+  .transform(d => ({
+    on_update_success: d.onUpdateSuccess,
+    on_update_failure: d.onUpdateFailure,
+  }));
+
+export const marshalEndpointCoreConfigSchema: z.ZodType = z
+  .object({
+    servedEntities: z.array(z.lazy(() => marshalServedModelSchema)).optional(),
+    servedModels: z.array(z.lazy(() => marshalServedModelSchema)).optional(),
+    trafficConfig: z.lazy(() => marshalTrafficConfigSchema).optional(),
+    autoCaptureConfig: z.lazy(() => marshalAutoCaptureConfigSchema).optional(),
+  })
+  .transform(d => ({
+    served_entities: d.servedEntities,
+    served_models: d.servedModels,
+    traffic_config: d.trafficConfig,
+    auto_capture_config: d.autoCaptureConfig,
+  }));
+
+export const marshalEndpointTagSchema: z.ZodType = z
+  .object({
+    key: z.string().optional(),
+    value: z.string().optional(),
+  })
+  .transform(d => ({
+    key: d.key,
+    value: d.value,
+  }));
+
+export const marshalExternalFunctionRequestSchema: z.ZodType = z
+  .object({
+    connectionName: z.string().optional(),
+    method: z.enum(ExternalFunctionRequest_HttpMethod).optional(),
+    path: z.string().optional(),
+    json: z.string().optional(),
+    headers: z.string().optional(),
+    params: z.string().optional(),
+    subDomain: z.string().optional(),
+  })
+  .transform(d => ({
+    connection_name: d.connectionName,
+    method: d.method,
+    path: d.path,
+    json: d.json,
+    headers: d.headers,
+    params: d.params,
+    sub_domain: d.subDomain,
+  }));
+
+export const marshalExternalModelSchema: z.ZodType = z
+  .object({
+    provider: z.string().optional(),
+    name: z.string().optional(),
+    task: z.string().optional(),
+    config: z
+      .discriminatedUnion('$case', [
+        z.object({
+          $case: z.literal('ai21labsConfig'),
+          ai21labsConfig: z.lazy(() => marshalAi21LabsConfigSchema),
+        }),
+        z.object({
+          $case: z.literal('anthropicConfig'),
+          anthropicConfig: z.lazy(() => marshalAnthropicConfigSchema),
+        }),
+        z.object({
+          $case: z.literal('amazonBedrockConfig'),
+          amazonBedrockConfig: z.lazy(() => marshalAmazonBedrockConfigSchema),
+        }),
+        z.object({
+          $case: z.literal('cohereConfig'),
+          cohereConfig: z.lazy(() => marshalCohereConfigSchema),
+        }),
+        z.object({
+          $case: z.literal('googleCloudVertexAiConfig'),
+          googleCloudVertexAiConfig: z.lazy(
+            () => marshalGoogleCloudVertexAiConfigSchema
+          ),
+        }),
+        z.object({
+          $case: z.literal('databricksModelServingConfig'),
+          databricksModelServingConfig: z.lazy(
+            () => marshalDatabricksModelServingConfigSchema
+          ),
+        }),
+        z.object({
+          $case: z.literal('openaiConfig'),
+          openaiConfig: z.lazy(() => marshalOpenAiConfigSchema),
+        }),
+        z.object({
+          $case: z.literal('palmConfig'),
+          palmConfig: z.lazy(() => marshalPaLmConfigSchema),
+        }),
+        z.object({
+          $case: z.literal('customProviderConfig'),
+          customProviderConfig: z.lazy(() => marshalCustomProviderConfigSchema),
+        }),
+      ])
+      .optional(),
+  })
+  .transform(d => ({
+    provider: d.provider,
+    name: d.name,
+    task: d.task,
+    ...(d.config?.$case === 'ai21labsConfig' && {
+      ai21labs_config: d.config.ai21labsConfig,
+    }),
+    ...(d.config?.$case === 'anthropicConfig' && {
+      anthropic_config: d.config.anthropicConfig,
+    }),
+    ...(d.config?.$case === 'amazonBedrockConfig' && {
+      amazon_bedrock_config: d.config.amazonBedrockConfig,
+    }),
+    ...(d.config?.$case === 'cohereConfig' && {
+      cohere_config: d.config.cohereConfig,
+    }),
+    ...(d.config?.$case === 'googleCloudVertexAiConfig' && {
+      google_cloud_vertex_ai_config: d.config.googleCloudVertexAiConfig,
+    }),
+    ...(d.config?.$case === 'databricksModelServingConfig' && {
+      databricks_model_serving_config: d.config.databricksModelServingConfig,
+    }),
+    ...(d.config?.$case === 'openaiConfig' && {
+      openai_config: d.config.openaiConfig,
+    }),
+    ...(d.config?.$case === 'palmConfig' && {palm_config: d.config.palmConfig}),
+    ...(d.config?.$case === 'customProviderConfig' && {
+      custom_provider_config: d.config.customProviderConfig,
+    }),
+  }));
+
+export const marshalFallbackConfigSchema: z.ZodType = z
+  .object({
+    enabled: z.boolean().optional(),
+  })
+  .transform(d => ({
+    enabled: d.enabled,
+  }));
+
+export const marshalFoundationModelSchema: z.ZodType = z
+  .object({
+    name: z.string().optional(),
+    displayName: z.string().optional(),
+    docs: z.string().optional(),
+    description: z.string().optional(),
+  })
+  .transform(d => ({
+    name: d.name,
+    display_name: d.displayName,
+    docs: d.docs,
+    description: d.description,
+  }));
+
+export const marshalGoogleCloudVertexAiConfigSchema: z.ZodType = z
+  .object({
+    privateKey: z.string().optional(),
+    projectId: z.string().optional(),
+    region: z.string().optional(),
+    privateKeyPlaintext: z.string().optional(),
+  })
+  .transform(d => ({
+    private_key: d.privateKey,
+    project_id: d.projectId,
+    region: d.region,
+    private_key_plaintext: d.privateKeyPlaintext,
+  }));
+
+export const marshalInferenceTableConfigSchema: z.ZodType = z
+  .object({
+    catalogName: z.string().optional(),
+    schemaName: z.string().optional(),
+    tableNamePrefix: z.string().optional(),
+    enabled: z.boolean().optional(),
+  })
+  .transform(d => ({
+    catalog_name: d.catalogName,
+    schema_name: d.schemaName,
+    table_name_prefix: d.tableNamePrefix,
+    enabled: d.enabled,
+  }));
+
+export const marshalOpenAiConfigSchema: z.ZodType = z
+  .object({
+    openaiApiKey: z.string().optional(),
+    openaiApiType: z.string().optional(),
+    openaiApiBase: z.string().optional(),
+    openaiApiVersion: z.string().optional(),
+    openaiDeploymentName: z.string().optional(),
+    openaiOrganization: z.string().optional(),
+    microsoftEntraTenantId: z.string().optional(),
+    microsoftEntraClientId: z.string().optional(),
+    microsoftEntraClientSecret: z.string().optional(),
+    openaiApiKeyPlaintext: z.string().optional(),
+    microsoftEntraClientSecretPlaintext: z.string().optional(),
+  })
+  .transform(d => ({
+    openai_api_key: d.openaiApiKey,
+    openai_api_type: d.openaiApiType,
+    openai_api_base: d.openaiApiBase,
+    openai_api_version: d.openaiApiVersion,
+    openai_deployment_name: d.openaiDeploymentName,
+    openai_organization: d.openaiOrganization,
+    microsoft_entra_tenant_id: d.microsoftEntraTenantId,
+    microsoft_entra_client_id: d.microsoftEntraClientId,
+    microsoft_entra_client_secret: d.microsoftEntraClientSecret,
+    openai_api_key_plaintext: d.openaiApiKeyPlaintext,
+    microsoft_entra_client_secret_plaintext:
+      d.microsoftEntraClientSecretPlaintext,
+  }));
+
+export const marshalPaLmConfigSchema: z.ZodType = z
+  .object({
+    palmApiKey: z.string().optional(),
+    palmApiKeyPlaintext: z.string().optional(),
+  })
+  .transform(d => ({
+    palm_api_key: d.palmApiKey,
+    palm_api_key_plaintext: d.palmApiKeyPlaintext,
+  }));
+
+export const marshalPatchInferenceEndpointTagsRequestSchema: z.ZodType = z
+  .object({
+    name: z.string().optional(),
+    addTags: z.array(z.lazy(() => marshalEndpointTagSchema)).optional(),
+    deleteTags: z.array(z.string()).optional(),
+  })
+  .transform(d => ({
+    name: d.name,
+    add_tags: d.addTags,
+    delete_tags: d.deleteTags,
+  }));
+
+export const marshalPayloadTableSchema: z.ZodType = z
+  .object({
+    name: z.string().optional(),
+    status: z.string().optional(),
+    statusMessage: z.string().optional(),
+  })
+  .transform(d => ({
+    name: d.name,
+    status: d.status,
+    status_message: d.statusMessage,
+  }));
+
+export const marshalPiiSettingsSchema: z.ZodType = z
+  .object({
+    behavior: z.enum(Behavior).optional(),
+  })
+  .transform(d => ({
+    behavior: d.behavior,
+  }));
+
+export const marshalPtEndpointCoreConfigSchema: z.ZodType = z
+  .object({
+    servedEntities: z
+      .array(z.lazy(() => marshalPtServedModelSchema))
+      .optional(),
+    trafficConfig: z.lazy(() => marshalTrafficConfigSchema).optional(),
+  })
+  .transform(d => ({
+    served_entities: d.servedEntities,
+    traffic_config: d.trafficConfig,
+  }));
+
+export const marshalPtServedModelSchema: z.ZodType = z
+  .object({
+    name: z.string().optional(),
+    entityName: z.string().optional(),
+    entityVersion: z.string().optional(),
+    provisionedModelUnits: z.bigint().optional(),
+    burstScalingEnabled: z.boolean().optional(),
+  })
+  .transform(d => ({
+    name: d.name,
+    entity_name: d.entityName,
+    entity_version: d.entityVersion,
+    provisioned_model_units: d.provisionedModelUnits,
+    burst_scaling_enabled: d.burstScalingEnabled,
+  }));
+
+export const marshalPutInferenceEndpointAiGatewayRequestSchema: z.ZodType = z
+  .object({
+    name: z.string().optional(),
+    usageTrackingConfig: z
+      .lazy(() => marshalUsageTrackingConfigSchema)
+      .optional(),
+    inferenceTableConfig: z
+      .lazy(() => marshalInferenceTableConfigSchema)
+      .optional(),
+    rateLimits: z
+      .array(z.lazy(() => marshalAiGatewayRateLimitSchema))
+      .optional(),
+    guardrails: z.lazy(() => marshalAiGuardrailsSchema).optional(),
+    fallbackConfig: z.lazy(() => marshalFallbackConfigSchema).optional(),
+  })
+  .transform(d => ({
+    name: d.name,
+    usage_tracking_config: d.usageTrackingConfig,
+    inference_table_config: d.inferenceTableConfig,
+    rate_limits: d.rateLimits,
+    guardrails: d.guardrails,
+    fallback_config: d.fallbackConfig,
+  }));
+
+export const marshalPutInferenceEndpointConfigRequestSchema: z.ZodType = z
+  .object({
+    name: z.string().optional(),
+    servedEntities: z.array(z.lazy(() => marshalServedModelSchema)).optional(),
+    servedModels: z.array(z.lazy(() => marshalServedModelSchema)).optional(),
+    trafficConfig: z.lazy(() => marshalTrafficConfigSchema).optional(),
+    autoCaptureConfig: z.lazy(() => marshalAutoCaptureConfigSchema).optional(),
+  })
+  .transform(d => ({
+    name: d.name,
+    served_entities: d.servedEntities,
+    served_models: d.servedModels,
+    traffic_config: d.trafficConfig,
+    auto_capture_config: d.autoCaptureConfig,
+  }));
+
+export const marshalPutInferenceEndpointRateLimitsRequestSchema: z.ZodType = z
+  .object({
+    name: z.string().optional(),
+    rateLimits: z.array(z.lazy(() => marshalRateLimitSchema)).optional(),
+  })
+  .transform(d => ({
+    name: d.name,
+    rate_limits: d.rateLimits,
+  }));
+
+export const marshalPutPtEndpointConfigRequestSchema: z.ZodType = z
+  .object({
+    name: z.string().optional(),
+    config: z.lazy(() => marshalPtEndpointCoreConfigSchema).optional(),
+  })
+  .transform(d => ({
+    name: d.name,
+    config: d.config,
+  }));
+
+export const marshalRateLimitSchema: z.ZodType = z
+  .object({
+    calls: z.bigint().optional(),
+    key: z.string().optional(),
+    renewalPeriod: z.string().optional(),
+  })
+  .transform(d => ({
+    calls: d.calls,
+    key: d.key,
+    renewal_period: d.renewalPeriod,
+  }));
+
+export const marshalRouteSchema: z.ZodType = z
+  .object({
+    servedModelName: z.string().optional(),
+    trafficPercentage: z.number().optional(),
+    servedEntityName: z.string().optional(),
+  })
+  .transform(d => ({
+    served_model_name: d.servedModelName,
+    traffic_percentage: d.trafficPercentage,
+    served_entity_name: d.servedEntityName,
+  }));
+
+export const marshalServedModelSchema: z.ZodType = z
+  .object({
+    name: z.string().optional(),
+    externalModel: z.lazy(() => marshalExternalModelSchema).optional(),
+    entityName: z.string().optional(),
+    entityVersion: z.string().optional(),
+    minProvisionedThroughput: z.number().optional(),
+    maxProvisionedThroughput: z.number().optional(),
+    minProvisionedConcurrency: z.number().optional(),
+    maxProvisionedConcurrency: z.number().optional(),
+    workloadSize: z.string().optional(),
+    provisionedModelUnits: z.bigint().optional(),
+    burstScalingEnabled: z.boolean().optional(),
+    scaleToZeroEnabled: z.boolean().optional(),
+    modelName: z.string().optional(),
+    modelVersion: z.string().optional(),
+    environmentVars: z.record(z.string(), z.string()).optional(),
+    instanceProfileArn: z.string().optional(),
+    foundationModel: z.lazy(() => marshalFoundationModelSchema).optional(),
+    state: z.lazy(() => marshalServedModelStateSchema).optional(),
+    creator: z.string().optional(),
+    creationTimestamp: z.bigint().optional(),
+  })
+  .transform(d => ({
+    name: d.name,
+    external_model: d.externalModel,
+    entity_name: d.entityName,
+    entity_version: d.entityVersion,
+    min_provisioned_throughput: d.minProvisionedThroughput,
+    max_provisioned_throughput: d.maxProvisionedThroughput,
+    min_provisioned_concurrency: d.minProvisionedConcurrency,
+    max_provisioned_concurrency: d.maxProvisionedConcurrency,
+    workload_size: d.workloadSize,
+    provisioned_model_units: d.provisionedModelUnits,
+    burst_scaling_enabled: d.burstScalingEnabled,
+    scale_to_zero_enabled: d.scaleToZeroEnabled,
+    model_name: d.modelName,
+    model_version: d.modelVersion,
+    environment_vars: d.environmentVars,
+    instance_profile_arn: d.instanceProfileArn,
+    foundation_model: d.foundationModel,
+    state: d.state,
+    creator: d.creator,
+    creation_timestamp: d.creationTimestamp,
+  }));
+
+export const marshalServedModelStateSchema: z.ZodType = z
+  .object({
+    deployment: z.enum(ServedModelDeploymentState).optional(),
+    deploymentStateMessage: z.string().optional(),
+  })
+  .transform(d => ({
+    deployment: d.deployment,
+    deployment_state_message: d.deploymentStateMessage,
+  }));
+
+export const marshalTrafficConfigSchema: z.ZodType = z
+  .object({
+    routes: z.array(z.lazy(() => marshalRouteSchema)).optional(),
+  })
+  .transform(d => ({
+    routes: d.routes,
+  }));
+
+export const marshalUpdateInferenceEndpointNotificationsRequestSchema: z.ZodType =
+  z
+    .object({
+      name: z.string().optional(),
+      emailNotifications: z
+        .lazy(() => marshalEmailNotificationsSchema)
+        .optional(),
+    })
+    .transform(d => ({
+      name: d.name,
+      email_notifications: d.emailNotifications,
+    }));
+
+export const marshalUsageTrackingConfigSchema: z.ZodType = z
+  .object({
+    enabled: z.boolean().optional(),
+  })
+  .transform(d => ({
+    enabled: d.enabled,
+  }));