# Guardrails > Guardrails - Use the endpoints described below to manage guardrails. Model guardrails are a set of > rules or machine learning models that ensure the output of a language model is safe, accurate, and > ethical. This Markdown file sits beside the HTML page at the same path (with a `.md` suffix). It summarizes the topic and lists links for tools and LLM context. Companion generated at `2026-05-06T18:17:09.715154+00:00` (UTC). ## Primary page - [Guardrails](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html): Full documentation for this topic (HTML). ## Sections on this page - [The list of resource tags](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#the-list-of-resource-tags): In-page section heading. - [Parameters](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#guardconfigurations_list-parameters): In-page section heading. - [Enumerated Values](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#enumerated-values): In-page section heading. - [Example responses](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#example-responses): In-page section heading. - [Responses](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#responses): In-page section heading. - [Create a guard configuration](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#create-a-guard-configuration): In-page section heading. - [Body parameter](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#body-parameter): In-page section heading. - [Parameters](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#guardconfigurations_create-parameters): In-page section heading. - [Example responses](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#example-responses_1): In-page section heading. - [Responses](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#responses_1): In-page section heading. - [Show the prediction environments in use](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#show-the-prediction-environments-in-use): In-page section heading. - [Parameters](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#guardconfigurationspredictionenvironmentsinuse_list-parameters): In-page section heading. - [Example responses](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#example-responses_2): In-page section heading. - [Responses](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#responses_2): In-page section heading. - [Apply moderation configuration](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#apply-moderation-configuration): In-page section heading. - [Body parameter](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#body-parameter_1): In-page section heading. - [Parameters](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#guardconfigurationstonewcustommodelversion_create-parameters): In-page section heading. - [Example responses](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#example-responses_3): In-page section heading. - [Responses](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#responses_3): In-page section heading. - [Delete a guard config by config ID](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#delete-a-guard-config-by-config-id): In-page section heading. - [Parameters](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#guardconfigurations_delete-parameters): In-page section heading. - [Responses](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#responses_4): In-page section heading. - [Retrieve info about a guard configuration by config ID](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#retrieve-info-about-a-guard-configuration-by-config-id): In-page section heading. - [Parameters](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#guardconfigurations_retrieve-parameters): In-page section heading. - [Example responses](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#example-responses_4): In-page section heading. - [Responses](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#responses_5): In-page section heading. - [Update a guard config by config ID](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#update-a-guard-config-by-config-id): In-page section heading. - [Body parameter](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#body-parameter_2): In-page section heading. - [Parameters](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#guardconfigurations_patch-parameters): In-page section heading. - [Example responses](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#example-responses_5): In-page section heading. - [Responses](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#responses_6): In-page section heading. - [List guard templates](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#list-guard-templates): In-page section heading. - [Parameters](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#guardtemplates_list-parameters): In-page section heading. - [Enumerated Values](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#enumerated-values_1): In-page section heading. - [Example responses](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#example-responses_6): In-page section heading. - [Responses](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#responses_7): In-page section heading. - [Retrieve information about a guard template by template ID](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#retrieve-information-about-a-guard-template-by-template-id): In-page section heading. - [Parameters](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#guardtemplates_retrieve-parameters): In-page section heading. - [Example responses](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#example-responses_7): In-page section heading. - [Responses](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#responses_8): In-page section heading. - [The list of supported LLMs](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#the-list-of-supported-llms): In-page section heading. - [Parameters](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#moderationsupportedllms_list-parameters): In-page section heading. - [Example responses](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#example-responses_8): In-page section heading. - [Responses](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#responses_9): In-page section heading. - [Get the overall moderation configuration](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#get-the-overall-moderation-configuration): In-page section heading. - [Parameters](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#overallmoderationconfiguration_list-parameters): In-page section heading. - [Enumerated Values](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#enumerated-values_2): In-page section heading. - [Example responses](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#example-responses_9): In-page section heading. - [Responses](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#responses_10): In-page section heading. - [Update the overall moderation configuration](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#update-the-overall-moderation-configuration): In-page section heading. - [Body parameter](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#body-parameter_3): In-page section heading. - [Parameters](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#overallmoderationconfiguration_patchmany-parameters): In-page section heading. - [Enumerated Values](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#enumerated-values_3): In-page section heading. - [Example responses](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#example-responses_10): In-page section heading. - [Responses](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#responses_11): In-page section heading. - [DeploymentAndGuardResponse](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#tocS_DeploymentAndGuardResponse): In-page section heading. - [Properties](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#properties): In-page section heading. - [GuardAdditionalConfig](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#tocS_GuardAdditionalConfig): In-page section heading. - [Properties](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#properties_1): In-page section heading. - [GuardConditionResponse](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#tocS_GuardConditionResponse): In-page section heading. - [Properties](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#properties_2): In-page section heading. - [Enumerated Values](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#enumerated-values_4): In-page section heading. - [GuardConfigurationConditionResponse](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#tocS_GuardConfigurationConditionResponse): In-page section heading. - [Properties](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#properties_3): In-page section heading. - [Enumerated Values](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#enumerated-values_5): In-page section heading. - [GuardConfigurationCostInfo](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#tocS_GuardConfigurationCostInfo): In-page section heading. - [Properties](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#properties_4): In-page section heading. - [GuardConfigurationCreate](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#tocS_GuardConfigurationCreate): In-page section heading. - [Properties](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#properties_5): In-page section heading. - [Enumerated Values](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#enumerated-values_6): In-page section heading. - [GuardConfigurationFullPost](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#tocS_GuardConfigurationFullPost): In-page section heading. - [Properties](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#properties_6): In-page section heading. - [Enumerated Values](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#enumerated-values_7): In-page section heading. - [GuardConfigurationInterventionResponse](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#tocS_GuardConfigurationInterventionResponse): In-page section heading. - [Properties](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#properties_7): In-page section heading. - [Enumerated Values](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#enumerated-values_8): In-page section heading. - [GuardConfigurationListResponse](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#tocS_GuardConfigurationListResponse): In-page section heading. - [Properties](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#properties_8): In-page section heading. - [GuardConfigurationNemoInfoResponse](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#tocS_GuardConfigurationNemoInfoResponse): In-page section heading. - [Properties](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#properties_9): In-page section heading. - [GuardConfigurationPayloadModelInfo](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#tocS_GuardConfigurationPayloadModelInfo): In-page section heading. - [Properties](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#properties_10): In-page section heading. - [Enumerated Values](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#enumerated-values_9): In-page section heading. - [GuardConfigurationPredictionEnvironmentsInUseListResponse](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#tocS_GuardConfigurationPredictionEnvironmentsInUseListResponse): In-page section heading. - [Properties](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#properties_11): In-page section heading. - [GuardConfigurationRetrieveResponse](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#tocS_GuardConfigurationRetrieveResponse): In-page section heading. - [Properties](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#properties_12): In-page section heading. - [Enumerated Values](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#enumerated-values_10): In-page section heading. - [GuardConfigurationToCustomModelResponse](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#tocS_GuardConfigurationToCustomModelResponse): In-page section heading. - [Properties](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#properties_13): In-page section heading. - [GuardConfigurationToCustomModelVersion](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#tocS_GuardConfigurationToCustomModelVersion): In-page section heading. - [Properties](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#properties_14): In-page section heading. - [GuardConfigurationUpdate](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#tocS_GuardConfigurationUpdate): In-page section heading. - [Properties](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#properties_15): In-page section heading. - [Enumerated Values](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#enumerated-values_11): In-page section heading. - [GuardInterventionResponse](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#tocS_GuardInterventionResponse): In-page section heading. - [Properties](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#properties_16): In-page section heading. - [Enumerated Values](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#enumerated-values_12): In-page section heading. - [GuardModelInfoResponse](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#tocS_GuardModelInfoResponse): In-page section heading. - [Properties](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#properties_17): In-page section heading. - [Enumerated Values](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#enumerated-values_13): In-page section heading. - [GuardNemoInfoResponse](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#tocS_GuardNemoInfoResponse): In-page section heading. - [Properties](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#properties_18): In-page section heading. - [GuardNemoLlmJudgeConfigResponse](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#tocS_GuardNemoLlmJudgeConfigResponse): In-page section heading. - [Properties](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#properties_19): In-page section heading. - [Enumerated Values](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#enumerated-values_14): In-page section heading. - [GuardNemoResponseRelevancyConfigResponse](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#tocS_GuardNemoResponseRelevancyConfigResponse): In-page section heading. - [Properties](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#properties_20): In-page section heading. - [GuardNemoTopicAdherenceConfigResponse](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#tocS_GuardNemoTopicAdherenceConfigResponse): In-page section heading. - [Properties](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#properties_21): In-page section heading. - [Enumerated Values](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#enumerated-values_15): In-page section heading. - [GuardTemplateListResponse](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#tocS_GuardTemplateListResponse): In-page section heading. - [Properties](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#properties_22): In-page section heading. - [GuardTemplateRetrieveResponse](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#tocS_GuardTemplateRetrieveResponse): In-page section heading. - [Properties](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#properties_23): In-page section heading. - [Enumerated Values](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#enumerated-values_16): In-page section heading. - [GuardToolCallInfo](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#tocS_GuardToolCallInfo): In-page section heading. - [Properties](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#properties_24): In-page section heading. - [Enumerated Values](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#enumerated-values_17): In-page section heading. - [OverallConfigUpdate](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#tocS_OverallConfigUpdate): In-page section heading. - [Properties](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#properties_25): In-page section heading. - [Enumerated Values](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#enumerated-values_18): In-page section heading. - [OverallModerationConfigurationResponse](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#tocS_OverallModerationConfigurationResponse): In-page section heading. - [Properties](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#properties_26): In-page section heading. - [Enumerated Values](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#enumerated-values_19): In-page section heading. - [OverallModerationConfigurationUpdate](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#tocS_OverallModerationConfigurationUpdate): In-page section heading. - [Properties](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#properties_27): In-page section heading. - [Enumerated Values](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#enumerated-values_20): In-page section heading. - [PredictionEnvironmentInUseResponse](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#tocS_PredictionEnvironmentInUseResponse): In-page section heading. - [Properties](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#properties_28): In-page section heading. - [SupportedLlmListResponse](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#tocS_SupportedLlmListResponse): In-page section heading. - [Properties](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#properties_29): In-page section heading. - [SupportedLlmResponse](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#tocS_SupportedLlmResponse): In-page section heading. - [Properties](https://docs.datarobot.com/en/docs/api/reference/public-api/guardrails.html#properties_30): In-page section heading. ## Related documentation - [Developer documentation](https://docs.datarobot.com/en/docs/api/index.html): Linked from this page. - [API reference](https://docs.datarobot.com/en/docs/api/reference/index.html): Linked from this page. - [REST API](https://docs.datarobot.com/en/docs/api/reference/public-api/index.html): Linked from this page. - [Secure and govern](https://docs.datarobot.com/en/docs/api/reference/public-api/tag-govern.html): Linked from this page. ## Documentation content Use the endpoints described below to manage guardrails. Model guardrails are a set of rules or machine learning models that ensure the output of a language model is safe, accurate, and ethical. ## The list of resource tags Operation path: `GET /api/v2/guardConfigurations/` Authentication requirements: `BearerAuth` The list of resource tags. ### Parameters | Name | In | Type | Required | Description | | --- | --- | --- | --- | --- | | offset | query | integer | false | This many results will be skipped. | | limit | query | integer | false | At most this many results are returned. | | entityId | query | string | true | Filter guard configurations by the given entity ID. | | entityType | query | string | true | The entity type of the given entity ID. | ### Enumerated Values | Parameter | Value | | --- | --- | | entityType | [customModel, customModelVersion, playground] | ### Example responses > 200 Response ``` { "properties": { "count": { "description": "The number of items returned on this page.", "type": "integer" }, "data": { "description": "The list of guard configurations.", "items": { "properties": { "additionalGuardConfig": { "description": "Additional configuration for the guard.", "properties": { "agentGuideline": { "description": "How to calculate agent guideline adherence for this guard.", "maxLength": 4096, "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "cost": { "description": "How to calculate cost information for this guard.", "properties": { "currency": { "description": "ISO 4217 Currency code for display.", "maxLength": 255, "type": "string" }, "inputPrice": { "description": "Cost per unit measure of tokens for input.", "minimum": 0, "type": "number" }, "inputUnit": { "description": "Number of tokens related to input price.", "minimum": 0, "type": "integer" }, "outputPrice": { "description": "Cost per unit measure of tokens for output.", "minimum": 0, "type": "number" }, "outputUnit": { "description": "Number of tokens related to output price.", "minimum": 0, "type": "integer" } }, "required": [ "currency", "inputPrice", "inputUnit", "outputPrice", "outputUnit" ], "type": "object", "x-versionadded": "v2.37" }, "toolCall": { "description": "How to calculate tool call metrics for this guard.", "properties": { "comparisonMetric": { "description": "How tool calls should be compared for metrics calculations.", "enum": [ "exactMatch", "doNotCompare" ], "type": "string" } }, "required": [ "comparisonMetric" ], "type": "object", "x-versionadded": "v2.37" } }, "type": "object", "x-versionadded": "v2.37" }, "allowedActions": { "description": "The actions this guard is allowed to take.", "items": { "enum": [ "block", "report", "replace" ], "type": "string" }, "maxItems": 10, "type": "array", "x-versionadded": "v2.36" }, "awsAccount": { "description": "The ID of the user credential containing an AWS account.", "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "awsModel": { "description": "AWS model.", "enum": [ "amazon-titan", "anthropic-claude-2", "anthropic-claude-3-haiku", "anthropic-claude-3-sonnet", "anthropic-claude-3-opus", "anthropic-claude-3.5-sonnet-v1", "anthropic-claude-3.5-sonnet-v2", "amazon-nova-lite", "amazon-nova-micro", "amazon-nova-pro" ], "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "awsRegion": { "description": "AWS model region.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "createdAt": { "description": "When the configuration was created.", "format": "date-time", "type": "string" }, "creatorId": { "description": "The ID of the user who created the guard configuration.", "type": [ "string", "null" ] }, "creatorName": { "description": "The name of the user who created the guard configuration.", "maxLength": 1000, "type": "string" }, "deploymentId": { "description": "The ID of the deployed model for model guards.", "type": [ "string", "null" ] }, "description": { "description": "The guard configuration description.", "maxLength": 4096, "type": "string" }, "entityId": { "description": "The ID of the custom model or playground for this guard.", "type": [ "string", "null" ] }, "entityType": { "description": "The type of the associated entity.", "enum": [ "customModel", "customModelVersion", "playground" ], "type": "string" }, "errorMessage": { "description": "Error message if the guard configuration is invalid.", "type": [ "string", "null" ] }, "googleModel": { "description": "Google model.", "enum": [ "chat-bison", "google-gemini-1.5-flash", "google-gemini-1.5-pro" ], "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "googleRegion": { "description": "Google model region.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "googleServiceAccount": { "description": "The ID of the user credential containing a Google service account.", "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "id": { "description": "The guard configuration object ID.", "type": "string" }, "intervention": { "description": "Intervention configuration for the guard.", "properties": { "action": { "description": "The action to take if conditions are met.", "enum": [ "block", "report", "replace" ], "type": "string" }, "allowedActions": { "description": "The actions this guard is allowed to take.", "items": { "enum": [ "block", "report", "replace" ], "type": "string" }, "maxItems": 10, "type": "array" }, "conditionLogic": { "default": "any", "description": "The action to take if conditions are met.", "enum": [ "any" ], "type": "string" }, "conditions": { "description": "The list of conditions to trigger intervention.", "items": { "description": "The condition to trigger intervention.", "properties": { "comparand": { "anyOf": [ { "type": "boolean" }, { "type": "number" }, { "type": "string" }, { "items": { "type": "string" }, "maxItems": 10, "type": "array" } ], "description": "The condition comparand (basis of comparison)." }, "comparator": { "description": "The condition comparator (operator).", "enum": [ "greaterThan", "lessThan", "equals", "notEquals", "is", "isNot", "matches", "doesNotMatch", "contains", "doesNotContain" ], "type": "string" } }, "required": [ "comparand", "comparator" ], "type": "object", "x-versionadded": "v2.35" }, "maxItems": 1, "type": "array" }, "message": { "description": "The message to use if the prompt or response is blocked.", "maxLength": 4096, "type": "string" }, "sendNotification": { "default": false, "description": "The notification event to create if intervention is triggered.", "type": "boolean" } }, "required": [ "action", "conditions", "message" ], "type": "object", "x-versionadded": "v2.35" }, "isAgentic": { "default": false, "description": "True if the guard is suitable for agentic workflows only.", "type": "boolean", "x-versionadded": "v2.37" }, "isValid": { "description": "Whether the guard is valid or not.", "type": "boolean" }, "llmGatewayModelId": { "description": "The LLM Gateway model ID to use as judge.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "llmType": { "description": "The type of LLM used by this guard.", "enum": [ "openAi", "azureOpenAi", "google", "amazon", "datarobot", "nim", "llmGateway" ], "type": [ "string", "null" ] }, "modelInfo": { "description": "The configuration info for guards using deployed models.", "properties": { "classNames": { "description": "The list of class names for multiclass models.", "items": { "description": "The class name.", "maxLength": 128, "type": "string" }, "maxItems": 100, "type": "array" }, "inputColumnName": { "description": "The input column name.", "maxLength": 255, "type": "string" }, "modelId": { "description": "The ID of the registered model for model guards.", "type": [ "string", "null" ] }, "modelName": { "default": "", "description": "The ID of the registered model for .model guards.", "maxLength": 255, "type": "string" }, "outputColumnName": { "description": "The output column name.", "maxLength": 255, "type": "string" }, "replacementTextColumnName": { "default": "", "description": "The name of the output column with replacement text. Required only if intervention.action is `replace`.", "maxLength": 255, "type": "string" }, "targetType": { "description": "The target type.", "enum": [ "Binary", "Regression", "Multiclass", "TextGeneration" ], "type": "string" } }, "required": [ "inputColumnName", "outputColumnName", "targetType" ], "type": "object", "x-versionadded": "v2.35" }, "name": { "description": "The guard configuration name.", "maxLength": 255, "type": "string" }, "nemoEvaluatorType": { "description": "Guard configuration \"NeMo Evaluator\" metric type", "enum": [ "llm_judge", "context_relevance", "response_groundedness", "topic_adherence", "agent_goal_accuracy", "response_relevancy", "faithfulness" ], "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "nemoInfo": { "description": "The configuration info for NeMo guards.", "properties": { "actions": { "description": "The NeMo guardrails actions file.", "maxLength": 4096, "type": "string" }, "blockedTerms": { "description": "The NeMo guardrails blocked terms list.", "maxLength": 4096, "type": "string" }, "credentialId": { "description": "The NeMo guardrails credential ID (deprecated; use \"openai_credential\").", "type": [ "string", "null" ] }, "llmPrompts": { "description": "The NeMo guardrails prompts.", "maxLength": 4096, "type": "string" }, "mainConfig": { "description": "The overall NeMo configuration YAML.", "maxLength": 4096, "type": "string" }, "railsConfig": { "description": "The NeMo guardrails configuration Colang.", "maxLength": 4096, "type": "string" } }, "required": [ "blockedTerms", "mainConfig", "railsConfig" ], "type": "object", "x-versionadded": "v2.35" }, "nemoLlmJudgeConfig": { "description": "The configuration for the LLM Judge metric.", "properties": { "customMetricDirectionality": { "description": "The custom metric directionality for the LLM judge.", "enum": [ "higherIsBetter", "lowerIsBetter" ], "type": "string", "x-versionadded": "v2.42" }, "scoreParsingRegex": { "description": "The regex to parse the score from the LLM judge response.", "maxLength": 1024, "type": "string" }, "systemPrompt": { "description": "The system prompt for the LLM judge.", "maxLength": 8192, "type": "string" }, "userPrompt": { "description": "The user prompt template for the LLM judge.", "maxLength": 8192, "type": "string" } }, "required": [ "customMetricDirectionality", "scoreParsingRegex", "systemPrompt", "userPrompt" ], "type": "object", "x-versionadded": "v2.42" }, "nemoResponseRelevancyConfig": { "description": "The configuration for the Response Relevancy metric.", "properties": { "embeddingDeploymentId": { "description": "The ID of the embedding model deployment.", "type": "string" } }, "required": [ "embeddingDeploymentId" ], "type": "object", "x-versionadded": "v2.42" }, "nemoTopicAdherenceConfig": { "description": "The configuration for the Topic Adherence metric.", "properties": { "metricMode": { "default": "f1", "description": "The metric calculation mode.", "enum": [ "f1", "recall", "precision" ], "type": "string" }, "referenceTopics": { "description": "The list of reference topics.", "items": { "description": "The reference topic.", "maxLength": 512, "type": "string" }, "maxItems": 100, "minItems": 1, "type": "array" } }, "required": [ "referenceTopics" ], "type": "object", "x-versionadded": "v2.42" }, "ootbType": { "description": "Guard template \"Out of the Box\" metric type", "enum": [ "token_count", "faithfulness", "rouge_1", "agent_goal_accuracy", "agent_goal_accuracy_with_reference", "cost", "task_adherence", "tool_call_accuracy", "agent_guideline_adherence", "agent_latency", "agent_tokens", "agent_cost", "agentGoalAccuracy", "agentGoalAccuracyWithReference", "taskAdherence", "toolCallAccuracy" ], "type": [ "string", "null" ] }, "openaiApiBase": { "description": "The Azure OpenAI API Base URL.", "maxLength": 255, "type": [ "string", "null" ] }, "openaiApiKey": { "description": "Deprecated; use openai_credential instead.", "maxLength": 255, "type": [ "string", "null" ] }, "openaiCredential": { "description": "The ID of the user credential containing an OpenAI token.", "type": [ "string", "null" ] }, "openaiDeploymentId": { "description": "The OpenAPI deployment ID.", "maxLength": 255, "type": [ "string", "null" ] }, "stages": { "description": "The stages where the guard is configured to run.", "items": { "enum": [ "prompt", "response" ], "type": "string" }, "maxItems": 16, "type": "array" }, "type": { "description": "The guard configuration type.", "enum": [ "guardModel", "nemo", "nemoEvaluator", "ootb", "pii", "userModel" ], "type": "string" } }, "required": [ "createdAt", "description", "entityId", "entityType", "id", "name", "stages", "type" ], "type": "object", "x-versionadded": "v2.35" }, "maxItems": 200, "type": "array" }, "next": { "description": "The URL of the next page (if null, there is no next page).", "format": "uri", "type": [ "string", "null" ] }, "previous": { "description": "The URL of the previous page (if null, there is no previous page).", "format": "uri", "type": [ "string", "null" ] }, "totalCount": { "description": "The total number of items across all pages.", "type": "integer" } }, "required": [ "data", "next", "previous", "totalCount" ], "type": "object", "x-versionadded": "v2.35" } ``` ### Responses | Status | Meaning | Description | Schema | | --- | --- | --- | --- | | 200 | OK | none | GuardConfigurationListResponse | | 400 | Bad Request | Request invalid, refer to messages for detail. | None | | 404 | Not Found | User permissions problem. | None | ## Create a guard configuration Operation path: `POST /api/v2/guardConfigurations/` Authentication requirements: `BearerAuth` Create a guard configuration. ### Body parameter ``` { "properties": { "additionalGuardConfig": { "description": "Additional configuration for the guard.", "properties": { "agentGuideline": { "description": "How to calculate agent guideline adherence for this guard.", "maxLength": 4096, "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "cost": { "description": "How to calculate cost information for this guard.", "properties": { "currency": { "description": "ISO 4217 Currency code for display.", "maxLength": 255, "type": "string" }, "inputPrice": { "description": "Cost per unit measure of tokens for input.", "minimum": 0, "type": "number" }, "inputUnit": { "description": "Number of tokens related to input price.", "minimum": 0, "type": "integer" }, "outputPrice": { "description": "Cost per unit measure of tokens for output.", "minimum": 0, "type": "number" }, "outputUnit": { "description": "Number of tokens related to output price.", "minimum": 0, "type": "integer" } }, "required": [ "currency", "inputPrice", "inputUnit", "outputPrice", "outputUnit" ], "type": "object", "x-versionadded": "v2.37" }, "toolCall": { "description": "How to calculate tool call metrics for this guard.", "properties": { "comparisonMetric": { "description": "How tool calls should be compared for metrics calculations.", "enum": [ "exactMatch", "doNotCompare" ], "type": "string" } }, "required": [ "comparisonMetric" ], "type": "object", "x-versionadded": "v2.37" } }, "type": "object", "x-versionadded": "v2.37" }, "allowedActions": { "description": "The actions this guard is allowed to take.", "items": { "enum": [ "block", "report", "replace" ], "type": "string" }, "maxItems": 10, "type": "array", "x-versionadded": "v2.36" }, "awsAccount": { "description": "The ID of the user credential containing an AWS account.", "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "awsModel": { "description": "AWS model.", "enum": [ "amazon-titan", "anthropic-claude-2", "anthropic-claude-3-haiku", "anthropic-claude-3-sonnet", "anthropic-claude-3-opus", "anthropic-claude-3.5-sonnet-v1", "anthropic-claude-3.5-sonnet-v2", "amazon-nova-lite", "amazon-nova-micro", "amazon-nova-pro" ], "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "awsRegion": { "description": "AWS model region.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "deploymentId": { "description": "The ID of the deployed model for model guards.", "type": [ "string", "null" ] }, "description": { "description": "Guard configuration description", "maxLength": 4096, "type": "string" }, "entityId": { "description": "The ID of the custom model or playground for this guard.", "type": "string" }, "entityType": { "description": "The type of the associated entity.", "enum": [ "customModel", "customModelVersion", "playground" ], "type": "string" }, "googleModel": { "description": "Google model.", "enum": [ "chat-bison", "google-gemini-1.5-flash", "google-gemini-1.5-pro" ], "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "googleRegion": { "description": "Google model region.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "googleServiceAccount": { "description": "The ID of the user credential containing a Google service account.", "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "intervention": { "description": "Intervention configuration for the guard.", "properties": { "action": { "description": "The action to take if conditions are met.", "enum": [ "block", "report", "replace" ], "type": "string" }, "allowedActions": { "description": "The actions this guard is allowed to take.", "items": { "enum": [ "block", "report", "replace" ], "type": "string" }, "maxItems": 10, "type": "array" }, "conditionLogic": { "default": "any", "description": "The action to take if conditions are met.", "enum": [ "any" ], "type": "string" }, "conditions": { "description": "The list of conditions to trigger intervention.", "items": { "description": "The condition to trigger intervention.", "properties": { "comparand": { "anyOf": [ { "type": "boolean" }, { "type": "number" }, { "type": "string" }, { "items": { "type": "string" }, "maxItems": 10, "type": "array" } ], "description": "The condition comparand (basis of comparison)." }, "comparator": { "description": "The condition comparator (operator).", "enum": [ "greaterThan", "lessThan", "equals", "notEquals", "is", "isNot", "matches", "doesNotMatch", "contains", "doesNotContain" ], "type": "string" } }, "required": [ "comparand", "comparator" ], "type": "object", "x-versionadded": "v2.35" }, "maxItems": 1, "type": "array" }, "message": { "description": "The message to use if the prompt or response is blocked.", "maxLength": 4096, "type": "string" }, "sendNotification": { "default": false, "description": "The notification event to create if intervention is triggered.", "type": "boolean" } }, "required": [ "action", "conditions", "message" ], "type": "object", "x-versionadded": "v2.35" }, "llmGatewayModelId": { "description": "The LLM Gateway model ID to use as judge.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "llmType": { "description": "The type of LLM used by this guard.", "enum": [ "openAi", "azureOpenAi", "google", "amazon", "datarobot", "nim", "llmGateway" ], "type": [ "string", "null" ] }, "modelInfo": { "description": "The configuration info for guards using deployed models.", "properties": { "classNames": { "description": "The list of class names for multiclass models.", "items": { "description": "The class name.", "maxLength": 128, "type": "string" }, "maxItems": 100, "type": "array" }, "inputColumnName": { "description": "The input column name.", "maxLength": 255, "type": "string" }, "modelId": { "description": "The ID of the registered model for model guards.", "type": [ "string", "null" ] }, "modelName": { "default": "", "description": "The ID of the registered model for .model guards.", "maxLength": 255, "type": "string" }, "outputColumnName": { "description": "The output column name.", "maxLength": 255, "type": "string" }, "replacementTextColumnName": { "default": "", "description": "The name of the output column with replacement text. Required only if intervention.action is `replace`.", "maxLength": 255, "type": "string" }, "targetType": { "description": "The target type.", "enum": [ "Binary", "Regression", "Multiclass", "TextGeneration" ], "type": [ "string", "null" ] } }, "required": [ "inputColumnName", "outputColumnName" ], "type": "object", "x-versionadded": "v2.35" }, "name": { "description": "Guard configuration name", "maxLength": 255, "type": "string" }, "nemoInfo": { "description": "The configuration info for NeMo guards.", "properties": { "actions": { "description": "The NeMo guardrails actions file.", "maxLength": 4096, "type": "string" }, "blockedTerms": { "description": "The NeMo guardrails blocked terms list.", "maxLength": 4096, "type": "string" }, "credentialId": { "description": "The NeMo guardrails credential ID (deprecated; use \"openai_credential\").", "type": [ "string", "null" ] }, "llmPrompts": { "description": "The NeMo guardrails prompts.", "maxLength": 4096, "type": "string" }, "mainConfig": { "description": "The overall NeMo configuration YAML.", "maxLength": 4096, "type": "string" }, "railsConfig": { "description": "The NeMo guardrails configuration Colang.", "maxLength": 4096, "type": "string" } }, "required": [ "blockedTerms", "mainConfig", "railsConfig" ], "type": "object", "x-versionadded": "v2.35" }, "nemoLlmJudgeConfig": { "description": "The configuration for the LLM Judge metric.", "properties": { "customMetricDirectionality": { "description": "The custom metric directionality for the LLM judge.", "enum": [ "higherIsBetter", "lowerIsBetter" ], "type": "string", "x-versionadded": "v2.42" }, "scoreParsingRegex": { "description": "The regex to parse the score from the LLM judge response.", "maxLength": 1024, "type": "string" }, "systemPrompt": { "description": "The system prompt for the LLM judge.", "maxLength": 8192, "type": "string" }, "userPrompt": { "description": "The user prompt template for the LLM judge.", "maxLength": 8192, "type": "string" } }, "required": [ "customMetricDirectionality", "scoreParsingRegex", "systemPrompt", "userPrompt" ], "type": "object", "x-versionadded": "v2.42" }, "nemoResponseRelevancyConfig": { "description": "The configuration for the Response Relevancy metric.", "properties": { "embeddingDeploymentId": { "description": "The ID of the embedding model deployment.", "type": "string" } }, "required": [ "embeddingDeploymentId" ], "type": "object", "x-versionadded": "v2.42" }, "nemoTopicAdherenceConfig": { "description": "The configuration for the Topic Adherence metric.", "properties": { "metricMode": { "default": "f1", "description": "The metric calculation mode.", "enum": [ "f1", "recall", "precision" ], "type": "string" }, "referenceTopics": { "description": "The list of reference topics.", "items": { "description": "The reference topic.", "maxLength": 512, "type": "string" }, "maxItems": 100, "minItems": 1, "type": "array" } }, "required": [ "referenceTopics" ], "type": "object", "x-versionadded": "v2.42" }, "openaiApiBase": { "description": "The Azure OpenAI API Base URL.", "maxLength": 255, "type": [ "string", "null" ] }, "openaiApiKey": { "description": "Deprecated; use openai_credential instead.", "maxLength": 255, "type": [ "string", "null" ] }, "openaiCredential": { "description": "The ID of the user credential containing an OpenAI token.", "type": [ "string", "null" ] }, "openaiDeploymentId": { "description": "The OpenAPI deployment ID.", "maxLength": 255, "type": [ "string", "null" ] }, "stages": { "description": "The stages where the guard can run.", "items": { "enum": [ "prompt", "response" ], "type": "string" }, "maxItems": 16, "type": "array" }, "templateId": { "description": "The ID of the template this guard is based on.", "type": "string" } }, "required": [ "entityId", "entityType", "name", "stages", "templateId" ], "type": "object", "x-versionadded": "v2.35" } ``` ### Parameters | Name | In | Type | Required | Description | | --- | --- | --- | --- | --- | | body | body | GuardConfigurationCreate | false | none | ### Example responses > 201 Response ``` { "properties": { "additionalGuardConfig": { "description": "Additional configuration for the guard.", "properties": { "agentGuideline": { "description": "How to calculate agent guideline adherence for this guard.", "maxLength": 4096, "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "cost": { "description": "How to calculate cost information for this guard.", "properties": { "currency": { "description": "ISO 4217 Currency code for display.", "maxLength": 255, "type": "string" }, "inputPrice": { "description": "Cost per unit measure of tokens for input.", "minimum": 0, "type": "number" }, "inputUnit": { "description": "Number of tokens related to input price.", "minimum": 0, "type": "integer" }, "outputPrice": { "description": "Cost per unit measure of tokens for output.", "minimum": 0, "type": "number" }, "outputUnit": { "description": "Number of tokens related to output price.", "minimum": 0, "type": "integer" } }, "required": [ "currency", "inputPrice", "inputUnit", "outputPrice", "outputUnit" ], "type": "object", "x-versionadded": "v2.37" }, "toolCall": { "description": "How to calculate tool call metrics for this guard.", "properties": { "comparisonMetric": { "description": "How tool calls should be compared for metrics calculations.", "enum": [ "exactMatch", "doNotCompare" ], "type": "string" } }, "required": [ "comparisonMetric" ], "type": "object", "x-versionadded": "v2.37" } }, "type": "object", "x-versionadded": "v2.37" }, "allowedActions": { "description": "The actions this guard is allowed to take.", "items": { "enum": [ "block", "report", "replace" ], "type": "string" }, "maxItems": 10, "type": "array", "x-versionadded": "v2.36" }, "awsAccount": { "description": "The ID of the user credential containing an AWS account.", "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "awsModel": { "description": "AWS model.", "enum": [ "amazon-titan", "anthropic-claude-2", "anthropic-claude-3-haiku", "anthropic-claude-3-sonnet", "anthropic-claude-3-opus", "anthropic-claude-3.5-sonnet-v1", "anthropic-claude-3.5-sonnet-v2", "amazon-nova-lite", "amazon-nova-micro", "amazon-nova-pro" ], "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "awsRegion": { "description": "AWS model region.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "createdAt": { "description": "When the configuration was created.", "format": "date-time", "type": "string" }, "creatorId": { "description": "The ID of the user who created the guard configuration.", "type": [ "string", "null" ] }, "creatorName": { "description": "The name of the user who created the guard configuration.", "maxLength": 1000, "type": "string" }, "deploymentId": { "description": "The ID of the deployed model for model guards.", "type": [ "string", "null" ] }, "description": { "description": "The guard configuration description.", "maxLength": 4096, "type": "string" }, "entityId": { "description": "The ID of the custom model or playground for this guard.", "type": [ "string", "null" ] }, "entityType": { "description": "The type of the associated entity.", "enum": [ "customModel", "customModelVersion", "playground" ], "type": "string" }, "errorMessage": { "description": "Error message if the guard configuration is invalid.", "type": [ "string", "null" ] }, "googleModel": { "description": "Google model.", "enum": [ "chat-bison", "google-gemini-1.5-flash", "google-gemini-1.5-pro" ], "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "googleRegion": { "description": "Google model region.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "googleServiceAccount": { "description": "The ID of the user credential containing a Google service account.", "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "id": { "description": "The guard configuration object ID.", "type": "string" }, "intervention": { "description": "Intervention configuration for the guard.", "properties": { "action": { "description": "The action to take if conditions are met.", "enum": [ "block", "report", "replace" ], "type": "string" }, "allowedActions": { "description": "The actions this guard is allowed to take.", "items": { "enum": [ "block", "report", "replace" ], "type": "string" }, "maxItems": 10, "type": "array" }, "conditionLogic": { "default": "any", "description": "The action to take if conditions are met.", "enum": [ "any" ], "type": "string" }, "conditions": { "description": "The list of conditions to trigger intervention.", "items": { "description": "The condition to trigger intervention.", "properties": { "comparand": { "anyOf": [ { "type": "boolean" }, { "type": "number" }, { "type": "string" }, { "items": { "type": "string" }, "maxItems": 10, "type": "array" } ], "description": "The condition comparand (basis of comparison)." }, "comparator": { "description": "The condition comparator (operator).", "enum": [ "greaterThan", "lessThan", "equals", "notEquals", "is", "isNot", "matches", "doesNotMatch", "contains", "doesNotContain" ], "type": "string" } }, "required": [ "comparand", "comparator" ], "type": "object", "x-versionadded": "v2.35" }, "maxItems": 1, "type": "array" }, "message": { "description": "The message to use if the prompt or response is blocked.", "maxLength": 4096, "type": "string" }, "sendNotification": { "default": false, "description": "The notification event to create if intervention is triggered.", "type": "boolean" } }, "required": [ "action", "conditions", "message" ], "type": "object", "x-versionadded": "v2.35" }, "isAgentic": { "default": false, "description": "True if the guard is suitable for agentic workflows only.", "type": "boolean", "x-versionadded": "v2.37" }, "isValid": { "description": "Whether the guard is valid or not.", "type": "boolean" }, "llmGatewayModelId": { "description": "The LLM Gateway model ID to use as judge.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "llmType": { "description": "The type of LLM used by this guard.", "enum": [ "openAi", "azureOpenAi", "google", "amazon", "datarobot", "nim", "llmGateway" ], "type": [ "string", "null" ] }, "modelInfo": { "description": "The configuration info for guards using deployed models.", "properties": { "classNames": { "description": "The list of class names for multiclass models.", "items": { "description": "The class name.", "maxLength": 128, "type": "string" }, "maxItems": 100, "type": "array" }, "inputColumnName": { "description": "The input column name.", "maxLength": 255, "type": "string" }, "modelId": { "description": "The ID of the registered model for model guards.", "type": [ "string", "null" ] }, "modelName": { "default": "", "description": "The ID of the registered model for .model guards.", "maxLength": 255, "type": "string" }, "outputColumnName": { "description": "The output column name.", "maxLength": 255, "type": "string" }, "replacementTextColumnName": { "default": "", "description": "The name of the output column with replacement text. Required only if intervention.action is `replace`.", "maxLength": 255, "type": "string" }, "targetType": { "description": "The target type.", "enum": [ "Binary", "Regression", "Multiclass", "TextGeneration" ], "type": "string" } }, "required": [ "inputColumnName", "outputColumnName", "targetType" ], "type": "object", "x-versionadded": "v2.35" }, "name": { "description": "The guard configuration name.", "maxLength": 255, "type": "string" }, "nemoEvaluatorType": { "description": "Guard configuration \"NeMo Evaluator\" metric type", "enum": [ "llm_judge", "context_relevance", "response_groundedness", "topic_adherence", "agent_goal_accuracy", "response_relevancy", "faithfulness" ], "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "nemoInfo": { "description": "The configuration info for NeMo guards.", "properties": { "actions": { "description": "The NeMo guardrails actions file.", "maxLength": 4096, "type": "string" }, "blockedTerms": { "description": "The NeMo guardrails blocked terms list.", "maxLength": 4096, "type": "string" }, "credentialId": { "description": "The NeMo guardrails credential ID (deprecated; use \"openai_credential\").", "type": [ "string", "null" ] }, "llmPrompts": { "description": "The NeMo guardrails prompts.", "maxLength": 4096, "type": "string" }, "mainConfig": { "description": "The overall NeMo configuration YAML.", "maxLength": 4096, "type": "string" }, "railsConfig": { "description": "The NeMo guardrails configuration Colang.", "maxLength": 4096, "type": "string" } }, "required": [ "blockedTerms", "mainConfig", "railsConfig" ], "type": "object", "x-versionadded": "v2.35" }, "nemoLlmJudgeConfig": { "description": "The configuration for the LLM Judge metric.", "properties": { "customMetricDirectionality": { "description": "The custom metric directionality for the LLM judge.", "enum": [ "higherIsBetter", "lowerIsBetter" ], "type": "string", "x-versionadded": "v2.42" }, "scoreParsingRegex": { "description": "The regex to parse the score from the LLM judge response.", "maxLength": 1024, "type": "string" }, "systemPrompt": { "description": "The system prompt for the LLM judge.", "maxLength": 8192, "type": "string" }, "userPrompt": { "description": "The user prompt template for the LLM judge.", "maxLength": 8192, "type": "string" } }, "required": [ "customMetricDirectionality", "scoreParsingRegex", "systemPrompt", "userPrompt" ], "type": "object", "x-versionadded": "v2.42" }, "nemoResponseRelevancyConfig": { "description": "The configuration for the Response Relevancy metric.", "properties": { "embeddingDeploymentId": { "description": "The ID of the embedding model deployment.", "type": "string" } }, "required": [ "embeddingDeploymentId" ], "type": "object", "x-versionadded": "v2.42" }, "nemoTopicAdherenceConfig": { "description": "The configuration for the Topic Adherence metric.", "properties": { "metricMode": { "default": "f1", "description": "The metric calculation mode.", "enum": [ "f1", "recall", "precision" ], "type": "string" }, "referenceTopics": { "description": "The list of reference topics.", "items": { "description": "The reference topic.", "maxLength": 512, "type": "string" }, "maxItems": 100, "minItems": 1, "type": "array" } }, "required": [ "referenceTopics" ], "type": "object", "x-versionadded": "v2.42" }, "ootbType": { "description": "Guard template \"Out of the Box\" metric type", "enum": [ "token_count", "faithfulness", "rouge_1", "agent_goal_accuracy", "agent_goal_accuracy_with_reference", "cost", "task_adherence", "tool_call_accuracy", "agent_guideline_adherence", "agent_latency", "agent_tokens", "agent_cost", "agentGoalAccuracy", "agentGoalAccuracyWithReference", "taskAdherence", "toolCallAccuracy" ], "type": [ "string", "null" ] }, "openaiApiBase": { "description": "The Azure OpenAI API Base URL.", "maxLength": 255, "type": [ "string", "null" ] }, "openaiApiKey": { "description": "Deprecated; use openai_credential instead.", "maxLength": 255, "type": [ "string", "null" ] }, "openaiCredential": { "description": "The ID of the user credential containing an OpenAI token.", "type": [ "string", "null" ] }, "openaiDeploymentId": { "description": "The OpenAPI deployment ID.", "maxLength": 255, "type": [ "string", "null" ] }, "stages": { "description": "The stages where the guard is configured to run.", "items": { "enum": [ "prompt", "response" ], "type": "string" }, "maxItems": 16, "type": "array" }, "type": { "description": "The guard configuration type.", "enum": [ "guardModel", "nemo", "nemoEvaluator", "ootb", "pii", "userModel" ], "type": "string" } }, "required": [ "createdAt", "description", "entityId", "entityType", "id", "name", "stages", "type" ], "type": "object", "x-versionadded": "v2.35" } ``` ### Responses | Status | Meaning | Description | Schema | | --- | --- | --- | --- | | 201 | Created | none | GuardConfigurationRetrieveResponse | | 404 | Not Found | Either the resource does not exist or the user does not have permission to create the configuration. | None | | 409 | Conflict | The proposed configuration name is already in use for the same entity. | None | ## Show the prediction environments in use Operation path: `GET /api/v2/guardConfigurations/predictionEnvironmentsInUse/` Authentication requirements: `BearerAuth` Show the prediction environments in use for moderation. ### Parameters | Name | In | Type | Required | Description | | --- | --- | --- | --- | --- | | offset | query | integer | false | This many results will be skipped. | | limit | query | integer | false | At most this many results are returned. | | customModelVersionId | query | string | true | Show prediction environment information for this custom model version. | ### Example responses > 200 Response ``` { "properties": { "count": { "description": "The number of items returned on this page.", "type": "integer" }, "data": { "description": "list of prediction environments in use for this custom model version.", "items": { "properties": { "id": { "description": "ID of prediction environment.", "type": "string" }, "name": { "description": "Name of prediction environment.", "type": "string" }, "usedBy": { "description": "Guards using this prediction environment.", "items": { "properties": { "configurationId": { "description": "ID of guard configuration.", "type": "string" }, "deploymentId": { "description": "ID of guard model deployment.", "type": "string" }, "name": { "description": "Name of guard configuration.", "type": "string" } }, "required": [ "configurationId", "deploymentId", "name" ], "type": "object", "x-versionadded": "v2.35" }, "maxItems": 32, "type": "array" } }, "required": [ "id", "name", "usedBy" ], "type": "object", "x-versionadded": "v2.35" }, "maxItems": 32, "type": "array" }, "next": { "description": "The URL of the next page (if null, there is no next page).", "format": "uri", "type": [ "string", "null" ] }, "previous": { "description": "The URL of the previous page (if null, there is no previous page).", "format": "uri", "type": [ "string", "null" ] }, "totalCount": { "description": "The total number of items across all pages.", "type": "integer" } }, "required": [ "data", "next", "previous", "totalCount" ], "type": "object", "x-versionadded": "v2.35" } ``` ### Responses | Status | Meaning | Description | Schema | | --- | --- | --- | --- | | 200 | OK | none | GuardConfigurationPredictionEnvironmentsInUseListResponse | | 404 | Not Found | Either the resource does not exist or the user does not have permission to view the configuration. | None | ## Apply moderation configuration Operation path: `POST /api/v2/guardConfigurations/toNewCustomModelVersion/` Authentication requirements: `BearerAuth` Apply moderation configuration to a new custom model version. ### Body parameter ``` { "properties": { "customModelId": { "description": "The ID of the custom model the user is working with.", "type": "string" }, "data": { "description": "The list of complete guard configurations to push.", "items": { "description": "Complete guard configuration to push", "properties": { "additionalGuardConfig": { "description": "Additional configuration for the guard.", "properties": { "agentGuideline": { "description": "How to calculate agent guideline adherence for this guard.", "maxLength": 4096, "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "cost": { "description": "How to calculate cost information for this guard.", "properties": { "currency": { "description": "ISO 4217 Currency code for display.", "maxLength": 255, "type": "string" }, "inputPrice": { "description": "Cost per unit measure of tokens for input.", "minimum": 0, "type": "number" }, "inputUnit": { "description": "Number of tokens related to input price.", "minimum": 0, "type": "integer" }, "outputPrice": { "description": "Cost per unit measure of tokens for output.", "minimum": 0, "type": "number" }, "outputUnit": { "description": "Number of tokens related to output price.", "minimum": 0, "type": "integer" } }, "required": [ "currency", "inputPrice", "inputUnit", "outputPrice", "outputUnit" ], "type": "object", "x-versionadded": "v2.37" }, "toolCall": { "description": "How to calculate tool call metrics for this guard.", "properties": { "comparisonMetric": { "description": "How tool calls should be compared for metrics calculations.", "enum": [ "exactMatch", "doNotCompare" ], "type": "string" } }, "required": [ "comparisonMetric" ], "type": "object", "x-versionadded": "v2.37" } }, "type": "object", "x-versionadded": "v2.37" }, "allowedActions": { "description": "The actions this guard is allowed to take.", "items": { "enum": [ "block", "report", "replace" ], "type": "string" }, "maxItems": 10, "type": "array", "x-versionadded": "v2.36" }, "awsAccount": { "description": "The ID of the user credential containing an AWS account.", "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "awsModel": { "description": "AWS model.", "enum": [ "amazon-titan", "anthropic-claude-2", "anthropic-claude-3-haiku", "anthropic-claude-3-sonnet", "anthropic-claude-3-opus", "anthropic-claude-3.5-sonnet-v1", "anthropic-claude-3.5-sonnet-v2", "amazon-nova-lite", "amazon-nova-micro", "amazon-nova-pro" ], "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "awsRegion": { "description": "AWS model region.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "deploymentId": { "description": "The ID of the deployed model for model guards.", "type": [ "string", "null" ] }, "description": { "description": "Guard configuration description", "maxLength": 4096, "type": "string" }, "errorMessage": { "description": "Error message if the guard configuration is invalid.", "type": [ "string", "null" ] }, "googleModel": { "description": "Google model.", "enum": [ "chat-bison", "google-gemini-1.5-flash", "google-gemini-1.5-pro" ], "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "googleRegion": { "description": "Google model region.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "googleServiceAccount": { "description": "The ID of the user credential containing a Google service account.", "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "intervention": { "description": "Intervention configuration for the guard.", "properties": { "action": { "description": "The action to take if conditions are met.", "enum": [ "block", "report", "replace" ], "type": "string" }, "allowedActions": { "description": "The actions this guard is allowed to take.", "items": { "enum": [ "block", "report", "replace" ], "type": "string" }, "maxItems": 10, "type": "array" }, "conditionLogic": { "default": "any", "description": "The action to take if conditions are met.", "enum": [ "any" ], "type": "string" }, "conditions": { "description": "The list of conditions to trigger intervention.", "items": { "description": "The condition to trigger intervention.", "properties": { "comparand": { "anyOf": [ { "type": "boolean" }, { "type": "number" }, { "type": "string" }, { "items": { "type": "string" }, "maxItems": 10, "type": "array" } ], "description": "The condition comparand (basis of comparison)." }, "comparator": { "description": "The condition comparator (operator).", "enum": [ "greaterThan", "lessThan", "equals", "notEquals", "is", "isNot", "matches", "doesNotMatch", "contains", "doesNotContain" ], "type": "string" } }, "required": [ "comparand", "comparator" ], "type": "object", "x-versionadded": "v2.35" }, "maxItems": 1, "type": "array" }, "message": { "description": "The message to use if the prompt or response is blocked.", "maxLength": 4096, "type": "string" }, "sendNotification": { "default": false, "description": "The notification event to create if intervention is triggered.", "type": "boolean" } }, "required": [ "action", "conditions", "message" ], "type": "object", "x-versionadded": "v2.35" }, "isAgentic": { "default": false, "description": "True if the guard is suitable for agentic workflows only.", "type": "boolean", "x-versionadded": "v2.37" }, "isValid": { "description": "Whether the guard is valid or not.", "type": "boolean" }, "llmGatewayModelId": { "description": "The LLM Gateway model ID to use as judge.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "llmType": { "description": "The type of LLM used by this guard.", "enum": [ "openAi", "azureOpenAi", "google", "amazon", "datarobot", "nim", "llmGateway" ], "type": [ "string", "null" ] }, "modelInfo": { "description": "The configuration info for guards using deployed models.", "properties": { "classNames": { "description": "The list of class names for multiclass models.", "items": { "description": "The class name.", "maxLength": 128, "type": "string" }, "maxItems": 100, "type": "array" }, "inputColumnName": { "description": "The input column name.", "maxLength": 255, "type": "string" }, "modelId": { "description": "The ID of the registered model for model guards.", "type": [ "string", "null" ] }, "modelName": { "default": "", "description": "The ID of the registered model for .model guards.", "maxLength": 255, "type": "string" }, "outputColumnName": { "description": "The output column name.", "maxLength": 255, "type": "string" }, "replacementTextColumnName": { "default": "", "description": "The name of the output column with replacement text. Required only if intervention.action is `replace`.", "maxLength": 255, "type": "string" }, "targetType": { "description": "The target type.", "enum": [ "Binary", "Regression", "Multiclass", "TextGeneration" ], "type": "string" } }, "required": [ "inputColumnName", "outputColumnName", "targetType" ], "type": "object", "x-versionadded": "v2.35" }, "name": { "description": "Guard configuration name", "maxLength": 255, "type": "string" }, "nemoEvaluatorType": { "description": "Guard configuration \"NeMo Evaluator\" metric type", "enum": [ "llm_judge", "context_relevance", "response_groundedness", "topic_adherence", "agent_goal_accuracy", "response_relevancy", "faithfulness" ], "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "nemoInfo": { "description": "The configuration info for NeMo guards.", "properties": { "actions": { "description": "The NeMo guardrails actions file.", "maxLength": 4096, "type": "string" }, "blockedTerms": { "description": "The NeMo guardrails blocked terms list.", "maxLength": 4096, "type": "string" }, "credentialId": { "description": "The NeMo guardrails credential ID (deprecated; use \"openai_credential\").", "type": [ "string", "null" ] }, "llmPrompts": { "description": "The NeMo guardrails prompts.", "maxLength": 4096, "type": "string" }, "mainConfig": { "description": "The overall NeMo configuration YAML.", "maxLength": 4096, "type": "string" }, "railsConfig": { "description": "The NeMo guardrails configuration Colang.", "maxLength": 4096, "type": "string" } }, "required": [ "blockedTerms", "mainConfig", "railsConfig" ], "type": "object", "x-versionadded": "v2.35" }, "nemoLlmJudgeConfig": { "description": "The configuration for the LLM Judge metric.", "properties": { "customMetricDirectionality": { "description": "The custom metric directionality for the LLM judge.", "enum": [ "higherIsBetter", "lowerIsBetter" ], "type": "string", "x-versionadded": "v2.42" }, "scoreParsingRegex": { "description": "The regex to parse the score from the LLM judge response.", "maxLength": 1024, "type": "string" }, "systemPrompt": { "description": "The system prompt for the LLM judge.", "maxLength": 8192, "type": "string" }, "userPrompt": { "description": "The user prompt template for the LLM judge.", "maxLength": 8192, "type": "string" } }, "required": [ "customMetricDirectionality", "scoreParsingRegex", "systemPrompt", "userPrompt" ], "type": "object", "x-versionadded": "v2.42" }, "nemoResponseRelevancyConfig": { "description": "The configuration for the Response Relevancy metric.", "properties": { "embeddingDeploymentId": { "description": "The ID of the embedding model deployment.", "type": "string" } }, "required": [ "embeddingDeploymentId" ], "type": "object", "x-versionadded": "v2.42" }, "nemoTopicAdherenceConfig": { "description": "The configuration for the Topic Adherence metric.", "properties": { "metricMode": { "default": "f1", "description": "The metric calculation mode.", "enum": [ "f1", "recall", "precision" ], "type": "string" }, "referenceTopics": { "description": "The list of reference topics.", "items": { "description": "The reference topic.", "maxLength": 512, "type": "string" }, "maxItems": 100, "minItems": 1, "type": "array" } }, "required": [ "referenceTopics" ], "type": "object", "x-versionadded": "v2.42" }, "ootbType": { "description": "Guard template \"Out of the Box\" metric type", "enum": [ "token_count", "faithfulness", "rouge_1", "agent_goal_accuracy", "agent_goal_accuracy_with_reference", "cost", "task_adherence", "tool_call_accuracy", "agent_guideline_adherence", "agent_latency", "agent_tokens", "agent_cost", "agentGoalAccuracy", "agentGoalAccuracyWithReference", "taskAdherence", "toolCallAccuracy" ], "type": [ "string", "null" ] }, "openaiApiBase": { "description": "The Azure OpenAI API Base URL.", "maxLength": 255, "type": [ "string", "null" ] }, "openaiApiKey": { "description": "Deprecated; use openai_credential instead.", "maxLength": 255, "type": [ "string", "null" ] }, "openaiCredential": { "description": "The ID of the user credential containing an OpenAI token.", "type": [ "string", "null" ] }, "openaiDeploymentId": { "description": "The OpenAI deployment ID.", "maxLength": 255, "type": [ "string", "null" ] }, "parameters": { "description": "Parameter list, not used, deprecated.", "items": { "maxLength": 1, "type": "string" }, "maxItems": 1, "type": "array" }, "stages": { "description": "The stages where the guard is configured to run.", "items": { "enum": [ "prompt", "response" ], "type": "string" }, "maxItems": 16, "type": "array" }, "type": { "description": "The guard configuration type.", "enum": [ "guardModel", "nemo", "nemoEvaluator", "ootb", "pii", "userModel" ], "type": "string" } }, "required": [ "description", "name", "stages", "type" ], "type": "object", "x-versionadded": "v2.35" }, "maxItems": 200, "type": "array" }, "overallConfig": { "description": "Overall moderation configuration to push (not specific to one guard)", "properties": { "nemoEvaluatorDeploymentId": { "description": "ID of NeMo Evaluator deployment to use for all NeMo Evaluator guards.", "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "timeoutAction": { "description": "Action to take if timeout occurs", "enum": [ "block", "score" ], "type": "string" }, "timeoutSec": { "description": "Timeout value in seconds for any guard", "minimum": 2, "type": "integer" } }, "required": [ "timeoutAction", "timeoutSec" ], "type": "object", "x-versionadded": "v2.35" } }, "required": [ "customModelId", "data" ], "type": "object", "x-versionadded": "v2.35" } ``` ### Parameters | Name | In | Type | Required | Description | | --- | --- | --- | --- | --- | | body | body | GuardConfigurationToCustomModelVersion | false | none | ### Example responses > 200 Response ``` { "properties": { "customModelVersionId": { "description": "The ID of the new custom model version created.", "type": "string" } }, "required": [ "customModelVersionId" ], "type": "object", "x-versionadded": "v2.35" } ``` ### Responses | Status | Meaning | Description | Schema | | --- | --- | --- | --- | | 200 | OK | none | GuardConfigurationToCustomModelResponse | | 404 | Not Found | Either the resource does not exist or the user does not have permission to create the configuration. | None | | 409 | Conflict | The destination custom model version is frozen. Create a new version to save configuration. | None | ## Delete a guard config by config ID Operation path: `DELETE /api/v2/guardConfigurations/{configId}/` Authentication requirements: `BearerAuth` Delete a guard config. ### Parameters | Name | In | Type | Required | Description | | --- | --- | --- | --- | --- | | configId | path | string | true | The ID of the configuration. | ### Responses | Status | Meaning | Description | Schema | | --- | --- | --- | --- | | 204 | No Content | none | None | | 404 | Not Found | Either the config does not exist or the user does not have permission to delete it. | None | ## Retrieve info about a guard configuration by config ID Operation path: `GET /api/v2/guardConfigurations/{configId}/` Authentication requirements: `BearerAuth` Retrieve info about a guard configuration. ### Parameters | Name | In | Type | Required | Description | | --- | --- | --- | --- | --- | | configId | path | string | true | The ID of the configuration. | ### Example responses > 200 Response ``` { "properties": { "additionalGuardConfig": { "description": "Additional configuration for the guard.", "properties": { "agentGuideline": { "description": "How to calculate agent guideline adherence for this guard.", "maxLength": 4096, "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "cost": { "description": "How to calculate cost information for this guard.", "properties": { "currency": { "description": "ISO 4217 Currency code for display.", "maxLength": 255, "type": "string" }, "inputPrice": { "description": "Cost per unit measure of tokens for input.", "minimum": 0, "type": "number" }, "inputUnit": { "description": "Number of tokens related to input price.", "minimum": 0, "type": "integer" }, "outputPrice": { "description": "Cost per unit measure of tokens for output.", "minimum": 0, "type": "number" }, "outputUnit": { "description": "Number of tokens related to output price.", "minimum": 0, "type": "integer" } }, "required": [ "currency", "inputPrice", "inputUnit", "outputPrice", "outputUnit" ], "type": "object", "x-versionadded": "v2.37" }, "toolCall": { "description": "How to calculate tool call metrics for this guard.", "properties": { "comparisonMetric": { "description": "How tool calls should be compared for metrics calculations.", "enum": [ "exactMatch", "doNotCompare" ], "type": "string" } }, "required": [ "comparisonMetric" ], "type": "object", "x-versionadded": "v2.37" } }, "type": "object", "x-versionadded": "v2.37" }, "allowedActions": { "description": "The actions this guard is allowed to take.", "items": { "enum": [ "block", "report", "replace" ], "type": "string" }, "maxItems": 10, "type": "array", "x-versionadded": "v2.36" }, "awsAccount": { "description": "The ID of the user credential containing an AWS account.", "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "awsModel": { "description": "AWS model.", "enum": [ "amazon-titan", "anthropic-claude-2", "anthropic-claude-3-haiku", "anthropic-claude-3-sonnet", "anthropic-claude-3-opus", "anthropic-claude-3.5-sonnet-v1", "anthropic-claude-3.5-sonnet-v2", "amazon-nova-lite", "amazon-nova-micro", "amazon-nova-pro" ], "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "awsRegion": { "description": "AWS model region.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "createdAt": { "description": "When the configuration was created.", "format": "date-time", "type": "string" }, "creatorId": { "description": "The ID of the user who created the guard configuration.", "type": [ "string", "null" ] }, "creatorName": { "description": "The name of the user who created the guard configuration.", "maxLength": 1000, "type": "string" }, "deploymentId": { "description": "The ID of the deployed model for model guards.", "type": [ "string", "null" ] }, "description": { "description": "The guard configuration description.", "maxLength": 4096, "type": "string" }, "entityId": { "description": "The ID of the custom model or playground for this guard.", "type": [ "string", "null" ] }, "entityType": { "description": "The type of the associated entity.", "enum": [ "customModel", "customModelVersion", "playground" ], "type": "string" }, "errorMessage": { "description": "Error message if the guard configuration is invalid.", "type": [ "string", "null" ] }, "googleModel": { "description": "Google model.", "enum": [ "chat-bison", "google-gemini-1.5-flash", "google-gemini-1.5-pro" ], "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "googleRegion": { "description": "Google model region.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "googleServiceAccount": { "description": "The ID of the user credential containing a Google service account.", "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "id": { "description": "The guard configuration object ID.", "type": "string" }, "intervention": { "description": "Intervention configuration for the guard.", "properties": { "action": { "description": "The action to take if conditions are met.", "enum": [ "block", "report", "replace" ], "type": "string" }, "allowedActions": { "description": "The actions this guard is allowed to take.", "items": { "enum": [ "block", "report", "replace" ], "type": "string" }, "maxItems": 10, "type": "array" }, "conditionLogic": { "default": "any", "description": "The action to take if conditions are met.", "enum": [ "any" ], "type": "string" }, "conditions": { "description": "The list of conditions to trigger intervention.", "items": { "description": "The condition to trigger intervention.", "properties": { "comparand": { "anyOf": [ { "type": "boolean" }, { "type": "number" }, { "type": "string" }, { "items": { "type": "string" }, "maxItems": 10, "type": "array" } ], "description": "The condition comparand (basis of comparison)." }, "comparator": { "description": "The condition comparator (operator).", "enum": [ "greaterThan", "lessThan", "equals", "notEquals", "is", "isNot", "matches", "doesNotMatch", "contains", "doesNotContain" ], "type": "string" } }, "required": [ "comparand", "comparator" ], "type": "object", "x-versionadded": "v2.35" }, "maxItems": 1, "type": "array" }, "message": { "description": "The message to use if the prompt or response is blocked.", "maxLength": 4096, "type": "string" }, "sendNotification": { "default": false, "description": "The notification event to create if intervention is triggered.", "type": "boolean" } }, "required": [ "action", "conditions", "message" ], "type": "object", "x-versionadded": "v2.35" }, "isAgentic": { "default": false, "description": "True if the guard is suitable for agentic workflows only.", "type": "boolean", "x-versionadded": "v2.37" }, "isValid": { "description": "Whether the guard is valid or not.", "type": "boolean" }, "llmGatewayModelId": { "description": "The LLM Gateway model ID to use as judge.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "llmType": { "description": "The type of LLM used by this guard.", "enum": [ "openAi", "azureOpenAi", "google", "amazon", "datarobot", "nim", "llmGateway" ], "type": [ "string", "null" ] }, "modelInfo": { "description": "The configuration info for guards using deployed models.", "properties": { "classNames": { "description": "The list of class names for multiclass models.", "items": { "description": "The class name.", "maxLength": 128, "type": "string" }, "maxItems": 100, "type": "array" }, "inputColumnName": { "description": "The input column name.", "maxLength": 255, "type": "string" }, "modelId": { "description": "The ID of the registered model for model guards.", "type": [ "string", "null" ] }, "modelName": { "default": "", "description": "The ID of the registered model for .model guards.", "maxLength": 255, "type": "string" }, "outputColumnName": { "description": "The output column name.", "maxLength": 255, "type": "string" }, "replacementTextColumnName": { "default": "", "description": "The name of the output column with replacement text. Required only if intervention.action is `replace`.", "maxLength": 255, "type": "string" }, "targetType": { "description": "The target type.", "enum": [ "Binary", "Regression", "Multiclass", "TextGeneration" ], "type": "string" } }, "required": [ "inputColumnName", "outputColumnName", "targetType" ], "type": "object", "x-versionadded": "v2.35" }, "name": { "description": "The guard configuration name.", "maxLength": 255, "type": "string" }, "nemoEvaluatorType": { "description": "Guard configuration \"NeMo Evaluator\" metric type", "enum": [ "llm_judge", "context_relevance", "response_groundedness", "topic_adherence", "agent_goal_accuracy", "response_relevancy", "faithfulness" ], "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "nemoInfo": { "description": "The configuration info for NeMo guards.", "properties": { "actions": { "description": "The NeMo guardrails actions file.", "maxLength": 4096, "type": "string" }, "blockedTerms": { "description": "The NeMo guardrails blocked terms list.", "maxLength": 4096, "type": "string" }, "credentialId": { "description": "The NeMo guardrails credential ID (deprecated; use \"openai_credential\").", "type": [ "string", "null" ] }, "llmPrompts": { "description": "The NeMo guardrails prompts.", "maxLength": 4096, "type": "string" }, "mainConfig": { "description": "The overall NeMo configuration YAML.", "maxLength": 4096, "type": "string" }, "railsConfig": { "description": "The NeMo guardrails configuration Colang.", "maxLength": 4096, "type": "string" } }, "required": [ "blockedTerms", "mainConfig", "railsConfig" ], "type": "object", "x-versionadded": "v2.35" }, "nemoLlmJudgeConfig": { "description": "The configuration for the LLM Judge metric.", "properties": { "customMetricDirectionality": { "description": "The custom metric directionality for the LLM judge.", "enum": [ "higherIsBetter", "lowerIsBetter" ], "type": "string", "x-versionadded": "v2.42" }, "scoreParsingRegex": { "description": "The regex to parse the score from the LLM judge response.", "maxLength": 1024, "type": "string" }, "systemPrompt": { "description": "The system prompt for the LLM judge.", "maxLength": 8192, "type": "string" }, "userPrompt": { "description": "The user prompt template for the LLM judge.", "maxLength": 8192, "type": "string" } }, "required": [ "customMetricDirectionality", "scoreParsingRegex", "systemPrompt", "userPrompt" ], "type": "object", "x-versionadded": "v2.42" }, "nemoResponseRelevancyConfig": { "description": "The configuration for the Response Relevancy metric.", "properties": { "embeddingDeploymentId": { "description": "The ID of the embedding model deployment.", "type": "string" } }, "required": [ "embeddingDeploymentId" ], "type": "object", "x-versionadded": "v2.42" }, "nemoTopicAdherenceConfig": { "description": "The configuration for the Topic Adherence metric.", "properties": { "metricMode": { "default": "f1", "description": "The metric calculation mode.", "enum": [ "f1", "recall", "precision" ], "type": "string" }, "referenceTopics": { "description": "The list of reference topics.", "items": { "description": "The reference topic.", "maxLength": 512, "type": "string" }, "maxItems": 100, "minItems": 1, "type": "array" } }, "required": [ "referenceTopics" ], "type": "object", "x-versionadded": "v2.42" }, "ootbType": { "description": "Guard template \"Out of the Box\" metric type", "enum": [ "token_count", "faithfulness", "rouge_1", "agent_goal_accuracy", "agent_goal_accuracy_with_reference", "cost", "task_adherence", "tool_call_accuracy", "agent_guideline_adherence", "agent_latency", "agent_tokens", "agent_cost", "agentGoalAccuracy", "agentGoalAccuracyWithReference", "taskAdherence", "toolCallAccuracy" ], "type": [ "string", "null" ] }, "openaiApiBase": { "description": "The Azure OpenAI API Base URL.", "maxLength": 255, "type": [ "string", "null" ] }, "openaiApiKey": { "description": "Deprecated; use openai_credential instead.", "maxLength": 255, "type": [ "string", "null" ] }, "openaiCredential": { "description": "The ID of the user credential containing an OpenAI token.", "type": [ "string", "null" ] }, "openaiDeploymentId": { "description": "The OpenAPI deployment ID.", "maxLength": 255, "type": [ "string", "null" ] }, "stages": { "description": "The stages where the guard is configured to run.", "items": { "enum": [ "prompt", "response" ], "type": "string" }, "maxItems": 16, "type": "array" }, "type": { "description": "The guard configuration type.", "enum": [ "guardModel", "nemo", "nemoEvaluator", "ootb", "pii", "userModel" ], "type": "string" } }, "required": [ "createdAt", "description", "entityId", "entityType", "id", "name", "stages", "type" ], "type": "object", "x-versionadded": "v2.35" } ``` ### Responses | Status | Meaning | Description | Schema | | --- | --- | --- | --- | | 200 | OK | none | GuardConfigurationRetrieveResponse | | 404 | Not Found | Either the config does not exist or the user does not have permission to view it. | None | ## Update a guard config by config ID Operation path: `PATCH /api/v2/guardConfigurations/{configId}/` Authentication requirements: `BearerAuth` Update a guard config. ### Body parameter ``` { "properties": { "additionalGuardConfig": { "description": "Additional configuration for the guard.", "properties": { "agentGuideline": { "description": "How to calculate agent guideline adherence for this guard.", "maxLength": 4096, "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "cost": { "description": "How to calculate cost information for this guard.", "properties": { "currency": { "description": "ISO 4217 Currency code for display.", "maxLength": 255, "type": "string" }, "inputPrice": { "description": "Cost per unit measure of tokens for input.", "minimum": 0, "type": "number" }, "inputUnit": { "description": "Number of tokens related to input price.", "minimum": 0, "type": "integer" }, "outputPrice": { "description": "Cost per unit measure of tokens for output.", "minimum": 0, "type": "number" }, "outputUnit": { "description": "Number of tokens related to output price.", "minimum": 0, "type": "integer" } }, "required": [ "currency", "inputPrice", "inputUnit", "outputPrice", "outputUnit" ], "type": "object", "x-versionadded": "v2.37" }, "toolCall": { "description": "How to calculate tool call metrics for this guard.", "properties": { "comparisonMetric": { "description": "How tool calls should be compared for metrics calculations.", "enum": [ "exactMatch", "doNotCompare" ], "type": "string" } }, "required": [ "comparisonMetric" ], "type": "object", "x-versionadded": "v2.37" } }, "type": "object", "x-versionadded": "v2.37" }, "allowedActions": { "description": "The actions this guard is allowed to take.", "items": { "enum": [ "block", "report", "replace" ], "type": "string" }, "maxItems": 10, "type": "array", "x-versionadded": "v2.36" }, "awsAccount": { "description": "The ID of the user credential containing an AWS account.", "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "awsModel": { "description": "AWS model.", "enum": [ "amazon-titan", "anthropic-claude-2", "anthropic-claude-3-haiku", "anthropic-claude-3-sonnet", "anthropic-claude-3-opus", "anthropic-claude-3.5-sonnet-v1", "anthropic-claude-3.5-sonnet-v2", "amazon-nova-lite", "amazon-nova-micro", "amazon-nova-pro" ], "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "awsRegion": { "description": "AWS model region.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "deploymentId": { "description": "The ID of the deployed model for model guards.", "type": [ "string", "null" ] }, "description": { "description": "Guard configuration description", "maxLength": 4096, "type": "string" }, "googleModel": { "description": "Google model.", "enum": [ "chat-bison", "google-gemini-1.5-flash", "google-gemini-1.5-pro" ], "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "googleRegion": { "description": "Google model region.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "googleServiceAccount": { "description": "The ID of the user credential containing a Google service account.", "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "intervention": { "description": "Intervention configuration for the guard.", "properties": { "action": { "description": "The action to take if conditions are met.", "enum": [ "block", "report", "replace" ], "type": "string" }, "allowedActions": { "description": "The actions this guard is allowed to take.", "items": { "enum": [ "block", "report", "replace" ], "type": "string" }, "maxItems": 10, "type": "array" }, "conditionLogic": { "default": "any", "description": "The action to take if conditions are met.", "enum": [ "any" ], "type": "string" }, "conditions": { "description": "The list of conditions to trigger intervention.", "items": { "description": "The condition to trigger intervention.", "properties": { "comparand": { "anyOf": [ { "type": "boolean" }, { "type": "number" }, { "type": "string" }, { "items": { "type": "string" }, "maxItems": 10, "type": "array" } ], "description": "The condition comparand (basis of comparison)." }, "comparator": { "description": "The condition comparator (operator).", "enum": [ "greaterThan", "lessThan", "equals", "notEquals", "is", "isNot", "matches", "doesNotMatch", "contains", "doesNotContain" ], "type": "string" } }, "required": [ "comparand", "comparator" ], "type": "object", "x-versionadded": "v2.35" }, "maxItems": 1, "type": "array" }, "message": { "description": "The message to use if the prompt or response is blocked.", "maxLength": 4096, "type": "string" }, "sendNotification": { "default": false, "description": "The notification event to create if intervention is triggered.", "type": "boolean" } }, "required": [ "action", "conditions", "message" ], "type": "object", "x-versionadded": "v2.35" }, "llmGatewayModelId": { "description": "The LLM Gateway model ID to use as judge.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "llmType": { "description": "The type of LLM used by this guard.", "enum": [ "openAi", "azureOpenAi", "google", "amazon", "datarobot", "nim", "llmGateway" ], "type": [ "string", "null" ] }, "modelInfo": { "description": "The configuration info for guards using deployed models.", "properties": { "classNames": { "description": "The list of class names for multiclass models.", "items": { "description": "The class name.", "maxLength": 128, "type": "string" }, "maxItems": 100, "type": "array" }, "inputColumnName": { "description": "The input column name.", "maxLength": 255, "type": "string" }, "modelId": { "description": "The ID of the registered model for model guards.", "type": [ "string", "null" ] }, "modelName": { "default": "", "description": "The ID of the registered model for .model guards.", "maxLength": 255, "type": "string" }, "outputColumnName": { "description": "The output column name.", "maxLength": 255, "type": "string" }, "replacementTextColumnName": { "default": "", "description": "The name of the output column with replacement text. Required only if intervention.action is `replace`.", "maxLength": 255, "type": "string" }, "targetType": { "description": "The target type.", "enum": [ "Binary", "Regression", "Multiclass", "TextGeneration" ], "type": [ "string", "null" ] } }, "required": [ "inputColumnName", "outputColumnName" ], "type": "object", "x-versionadded": "v2.35" }, "name": { "description": "Guard configuration name", "maxLength": 255, "type": "string" }, "nemoInfo": { "description": "The configuration info for NeMo guards.", "properties": { "actions": { "description": "The NeMo guardrails actions file.", "maxLength": 4096, "type": "string" }, "blockedTerms": { "description": "The NeMo guardrails blocked terms list.", "maxLength": 4096, "type": "string" }, "credentialId": { "description": "The NeMo guardrails credential ID (deprecated; use \"openai_credential\").", "type": [ "string", "null" ] }, "llmPrompts": { "description": "The NeMo guardrails prompts.", "maxLength": 4096, "type": "string" }, "mainConfig": { "description": "The overall NeMo configuration YAML.", "maxLength": 4096, "type": "string" }, "railsConfig": { "description": "The NeMo guardrails configuration Colang.", "maxLength": 4096, "type": "string" } }, "required": [ "blockedTerms", "mainConfig", "railsConfig" ], "type": "object", "x-versionadded": "v2.35" }, "nemoLlmJudgeConfig": { "description": "The configuration for the LLM Judge metric.", "properties": { "customMetricDirectionality": { "description": "The custom metric directionality for the LLM judge.", "enum": [ "higherIsBetter", "lowerIsBetter" ], "type": "string", "x-versionadded": "v2.42" }, "scoreParsingRegex": { "description": "The regex to parse the score from the LLM judge response.", "maxLength": 1024, "type": "string" }, "systemPrompt": { "description": "The system prompt for the LLM judge.", "maxLength": 8192, "type": "string" }, "userPrompt": { "description": "The user prompt template for the LLM judge.", "maxLength": 8192, "type": "string" } }, "required": [ "customMetricDirectionality", "scoreParsingRegex", "systemPrompt", "userPrompt" ], "type": "object", "x-versionadded": "v2.42" }, "nemoResponseRelevancyConfig": { "description": "The configuration for the Response Relevancy metric.", "properties": { "embeddingDeploymentId": { "description": "The ID of the embedding model deployment.", "type": "string" } }, "required": [ "embeddingDeploymentId" ], "type": "object", "x-versionadded": "v2.42" }, "nemoTopicAdherenceConfig": { "description": "The configuration for the Topic Adherence metric.", "properties": { "metricMode": { "default": "f1", "description": "The metric calculation mode.", "enum": [ "f1", "recall", "precision" ], "type": "string" }, "referenceTopics": { "description": "The list of reference topics.", "items": { "description": "The reference topic.", "maxLength": 512, "type": "string" }, "maxItems": 100, "minItems": 1, "type": "array" } }, "required": [ "referenceTopics" ], "type": "object", "x-versionadded": "v2.42" }, "openaiApiBase": { "description": "The Azure OpenAI API Base URL.", "maxLength": 255, "type": [ "string", "null" ] }, "openaiApiKey": { "description": "Deprecated; use openai_credential instead.", "maxLength": 255, "type": [ "string", "null" ] }, "openaiCredential": { "description": "The ID of the user credential containing an OpenAI token.", "type": [ "string", "null" ] }, "openaiDeploymentId": { "description": "The OpenAPI deployment ID.", "maxLength": 255, "type": [ "string", "null" ] } }, "type": "object", "x-versionadded": "v2.35" } ``` ### Parameters | Name | In | Type | Required | Description | | --- | --- | --- | --- | --- | | configId | path | string | true | The ID of the configuration. | | body | body | GuardConfigurationUpdate | false | none | ### Example responses > 200 Response ``` { "properties": { "additionalGuardConfig": { "description": "Additional configuration for the guard.", "properties": { "agentGuideline": { "description": "How to calculate agent guideline adherence for this guard.", "maxLength": 4096, "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "cost": { "description": "How to calculate cost information for this guard.", "properties": { "currency": { "description": "ISO 4217 Currency code for display.", "maxLength": 255, "type": "string" }, "inputPrice": { "description": "Cost per unit measure of tokens for input.", "minimum": 0, "type": "number" }, "inputUnit": { "description": "Number of tokens related to input price.", "minimum": 0, "type": "integer" }, "outputPrice": { "description": "Cost per unit measure of tokens for output.", "minimum": 0, "type": "number" }, "outputUnit": { "description": "Number of tokens related to output price.", "minimum": 0, "type": "integer" } }, "required": [ "currency", "inputPrice", "inputUnit", "outputPrice", "outputUnit" ], "type": "object", "x-versionadded": "v2.37" }, "toolCall": { "description": "How to calculate tool call metrics for this guard.", "properties": { "comparisonMetric": { "description": "How tool calls should be compared for metrics calculations.", "enum": [ "exactMatch", "doNotCompare" ], "type": "string" } }, "required": [ "comparisonMetric" ], "type": "object", "x-versionadded": "v2.37" } }, "type": "object", "x-versionadded": "v2.37" }, "allowedActions": { "description": "The actions this guard is allowed to take.", "items": { "enum": [ "block", "report", "replace" ], "type": "string" }, "maxItems": 10, "type": "array", "x-versionadded": "v2.36" }, "awsAccount": { "description": "The ID of the user credential containing an AWS account.", "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "awsModel": { "description": "AWS model.", "enum": [ "amazon-titan", "anthropic-claude-2", "anthropic-claude-3-haiku", "anthropic-claude-3-sonnet", "anthropic-claude-3-opus", "anthropic-claude-3.5-sonnet-v1", "anthropic-claude-3.5-sonnet-v2", "amazon-nova-lite", "amazon-nova-micro", "amazon-nova-pro" ], "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "awsRegion": { "description": "AWS model region.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "createdAt": { "description": "When the configuration was created.", "format": "date-time", "type": "string" }, "creatorId": { "description": "The ID of the user who created the guard configuration.", "type": [ "string", "null" ] }, "creatorName": { "description": "The name of the user who created the guard configuration.", "maxLength": 1000, "type": "string" }, "deploymentId": { "description": "The ID of the deployed model for model guards.", "type": [ "string", "null" ] }, "description": { "description": "The guard configuration description.", "maxLength": 4096, "type": "string" }, "entityId": { "description": "The ID of the custom model or playground for this guard.", "type": [ "string", "null" ] }, "entityType": { "description": "The type of the associated entity.", "enum": [ "customModel", "customModelVersion", "playground" ], "type": "string" }, "errorMessage": { "description": "Error message if the guard configuration is invalid.", "type": [ "string", "null" ] }, "googleModel": { "description": "Google model.", "enum": [ "chat-bison", "google-gemini-1.5-flash", "google-gemini-1.5-pro" ], "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "googleRegion": { "description": "Google model region.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "googleServiceAccount": { "description": "The ID of the user credential containing a Google service account.", "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "id": { "description": "The guard configuration object ID.", "type": "string" }, "intervention": { "description": "Intervention configuration for the guard.", "properties": { "action": { "description": "The action to take if conditions are met.", "enum": [ "block", "report", "replace" ], "type": "string" }, "allowedActions": { "description": "The actions this guard is allowed to take.", "items": { "enum": [ "block", "report", "replace" ], "type": "string" }, "maxItems": 10, "type": "array" }, "conditionLogic": { "default": "any", "description": "The action to take if conditions are met.", "enum": [ "any" ], "type": "string" }, "conditions": { "description": "The list of conditions to trigger intervention.", "items": { "description": "The condition to trigger intervention.", "properties": { "comparand": { "anyOf": [ { "type": "boolean" }, { "type": "number" }, { "type": "string" }, { "items": { "type": "string" }, "maxItems": 10, "type": "array" } ], "description": "The condition comparand (basis of comparison)." }, "comparator": { "description": "The condition comparator (operator).", "enum": [ "greaterThan", "lessThan", "equals", "notEquals", "is", "isNot", "matches", "doesNotMatch", "contains", "doesNotContain" ], "type": "string" } }, "required": [ "comparand", "comparator" ], "type": "object", "x-versionadded": "v2.35" }, "maxItems": 1, "type": "array" }, "message": { "description": "The message to use if the prompt or response is blocked.", "maxLength": 4096, "type": "string" }, "sendNotification": { "default": false, "description": "The notification event to create if intervention is triggered.", "type": "boolean" } }, "required": [ "action", "conditions", "message" ], "type": "object", "x-versionadded": "v2.35" }, "isAgentic": { "default": false, "description": "True if the guard is suitable for agentic workflows only.", "type": "boolean", "x-versionadded": "v2.37" }, "isValid": { "description": "Whether the guard is valid or not.", "type": "boolean" }, "llmGatewayModelId": { "description": "The LLM Gateway model ID to use as judge.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "llmType": { "description": "The type of LLM used by this guard.", "enum": [ "openAi", "azureOpenAi", "google", "amazon", "datarobot", "nim", "llmGateway" ], "type": [ "string", "null" ] }, "modelInfo": { "description": "The configuration info for guards using deployed models.", "properties": { "classNames": { "description": "The list of class names for multiclass models.", "items": { "description": "The class name.", "maxLength": 128, "type": "string" }, "maxItems": 100, "type": "array" }, "inputColumnName": { "description": "The input column name.", "maxLength": 255, "type": "string" }, "modelId": { "description": "The ID of the registered model for model guards.", "type": [ "string", "null" ] }, "modelName": { "default": "", "description": "The ID of the registered model for .model guards.", "maxLength": 255, "type": "string" }, "outputColumnName": { "description": "The output column name.", "maxLength": 255, "type": "string" }, "replacementTextColumnName": { "default": "", "description": "The name of the output column with replacement text. Required only if intervention.action is `replace`.", "maxLength": 255, "type": "string" }, "targetType": { "description": "The target type.", "enum": [ "Binary", "Regression", "Multiclass", "TextGeneration" ], "type": "string" } }, "required": [ "inputColumnName", "outputColumnName", "targetType" ], "type": "object", "x-versionadded": "v2.35" }, "name": { "description": "The guard configuration name.", "maxLength": 255, "type": "string" }, "nemoEvaluatorType": { "description": "Guard configuration \"NeMo Evaluator\" metric type", "enum": [ "llm_judge", "context_relevance", "response_groundedness", "topic_adherence", "agent_goal_accuracy", "response_relevancy", "faithfulness" ], "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "nemoInfo": { "description": "The configuration info for NeMo guards.", "properties": { "actions": { "description": "The NeMo guardrails actions file.", "maxLength": 4096, "type": "string" }, "blockedTerms": { "description": "The NeMo guardrails blocked terms list.", "maxLength": 4096, "type": "string" }, "credentialId": { "description": "The NeMo guardrails credential ID (deprecated; use \"openai_credential\").", "type": [ "string", "null" ] }, "llmPrompts": { "description": "The NeMo guardrails prompts.", "maxLength": 4096, "type": "string" }, "mainConfig": { "description": "The overall NeMo configuration YAML.", "maxLength": 4096, "type": "string" }, "railsConfig": { "description": "The NeMo guardrails configuration Colang.", "maxLength": 4096, "type": "string" } }, "required": [ "blockedTerms", "mainConfig", "railsConfig" ], "type": "object", "x-versionadded": "v2.35" }, "nemoLlmJudgeConfig": { "description": "The configuration for the LLM Judge metric.", "properties": { "customMetricDirectionality": { "description": "The custom metric directionality for the LLM judge.", "enum": [ "higherIsBetter", "lowerIsBetter" ], "type": "string", "x-versionadded": "v2.42" }, "scoreParsingRegex": { "description": "The regex to parse the score from the LLM judge response.", "maxLength": 1024, "type": "string" }, "systemPrompt": { "description": "The system prompt for the LLM judge.", "maxLength": 8192, "type": "string" }, "userPrompt": { "description": "The user prompt template for the LLM judge.", "maxLength": 8192, "type": "string" } }, "required": [ "customMetricDirectionality", "scoreParsingRegex", "systemPrompt", "userPrompt" ], "type": "object", "x-versionadded": "v2.42" }, "nemoResponseRelevancyConfig": { "description": "The configuration for the Response Relevancy metric.", "properties": { "embeddingDeploymentId": { "description": "The ID of the embedding model deployment.", "type": "string" } }, "required": [ "embeddingDeploymentId" ], "type": "object", "x-versionadded": "v2.42" }, "nemoTopicAdherenceConfig": { "description": "The configuration for the Topic Adherence metric.", "properties": { "metricMode": { "default": "f1", "description": "The metric calculation mode.", "enum": [ "f1", "recall", "precision" ], "type": "string" }, "referenceTopics": { "description": "The list of reference topics.", "items": { "description": "The reference topic.", "maxLength": 512, "type": "string" }, "maxItems": 100, "minItems": 1, "type": "array" } }, "required": [ "referenceTopics" ], "type": "object", "x-versionadded": "v2.42" }, "ootbType": { "description": "Guard template \"Out of the Box\" metric type", "enum": [ "token_count", "faithfulness", "rouge_1", "agent_goal_accuracy", "agent_goal_accuracy_with_reference", "cost", "task_adherence", "tool_call_accuracy", "agent_guideline_adherence", "agent_latency", "agent_tokens", "agent_cost", "agentGoalAccuracy", "agentGoalAccuracyWithReference", "taskAdherence", "toolCallAccuracy" ], "type": [ "string", "null" ] }, "openaiApiBase": { "description": "The Azure OpenAI API Base URL.", "maxLength": 255, "type": [ "string", "null" ] }, "openaiApiKey": { "description": "Deprecated; use openai_credential instead.", "maxLength": 255, "type": [ "string", "null" ] }, "openaiCredential": { "description": "The ID of the user credential containing an OpenAI token.", "type": [ "string", "null" ] }, "openaiDeploymentId": { "description": "The OpenAPI deployment ID.", "maxLength": 255, "type": [ "string", "null" ] }, "stages": { "description": "The stages where the guard is configured to run.", "items": { "enum": [ "prompt", "response" ], "type": "string" }, "maxItems": 16, "type": "array" }, "type": { "description": "The guard configuration type.", "enum": [ "guardModel", "nemo", "nemoEvaluator", "ootb", "pii", "userModel" ], "type": "string" } }, "required": [ "createdAt", "description", "entityId", "entityType", "id", "name", "stages", "type" ], "type": "object", "x-versionadded": "v2.35" } ``` ### Responses | Status | Meaning | Description | Schema | | --- | --- | --- | --- | | 200 | OK | none | GuardConfigurationRetrieveResponse | | 404 | Not Found | Either the resource does not exist or the user does not have permission to create the configuration. | None | | 409 | Conflict | The proposed configuration name is already in use for the same entity. | None | ## List guard templates Operation path: `GET /api/v2/guardTemplates/` Authentication requirements: `BearerAuth` List guard templates. ### Parameters | Name | In | Type | Required | Description | | --- | --- | --- | --- | --- | | offset | query | integer | false | This many results will be skipped. | | limit | query | integer | false | At most this many results are returned. | | includeAgentic | query | string | false | Include agentic and non-agentic templates in results (by default, agentic templates are not included). | | isAgentic | query | string | false | Limit results to agentic templates. If true, this overrides either includeAgentic setting. | | forPlayground | query | string | false | Filter for templates suitable for playground (exclude production-only templates). | | forProduction | query | string | false | Filter for templates suitable for production (exclude playground-only templates). | | name | query | string,null | false | Search for templates by name. | ### Enumerated Values | Parameter | Value | | --- | --- | | includeAgentic | [false, False, true, True] | | isAgentic | [false, False, true, True] | | forPlayground | [false, False, true, True] | | forProduction | [false, False, true, True] | ### Example responses > 200 Response ``` { "properties": { "count": { "description": "The number of items returned on this page.", "type": "integer" }, "data": { "description": "The list of guard templates.", "items": { "properties": { "allowedActions": { "description": "The actions this guard is allowed to take.", "items": { "enum": [ "block", "report", "replace" ], "type": "string" }, "maxItems": 10, "type": "array", "x-versionadded": "v2.36" }, "allowedStages": { "description": "The stages where the guard can run.", "items": { "enum": [ "prompt", "response" ], "type": "string" }, "maxItems": 16, "type": "array" }, "awsModel": { "description": "AWS model.", "enum": [ "amazon-titan", "anthropic-claude-2", "anthropic-claude-3-haiku", "anthropic-claude-3-sonnet", "anthropic-claude-3-opus", "anthropic-claude-3.5-sonnet-v1", "anthropic-claude-3.5-sonnet-v2", "amazon-nova-lite", "amazon-nova-micro", "amazon-nova-pro" ], "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "awsRegion": { "description": "AWS model region.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "createdAt": { "description": "When the template was created.", "format": "date-time", "type": "string" }, "creatorId": { "description": "The ID of the user who created the guard template.", "type": [ "string", "null" ] }, "creatorName": { "description": "The name of the user who created the guard template.", "maxLength": 1000, "type": "string" }, "description": { "description": "The guard template description.", "maxLength": 4096, "type": "string" }, "errorMessage": { "description": "Error message if the guard configuration is invalid.", "type": [ "string", "null" ] }, "googleModel": { "description": "Google model.", "enum": [ "chat-bison", "google-gemini-1.5-flash", "google-gemini-1.5-pro" ], "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "googleRegion": { "description": "Google model region.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "id": { "description": "The guard template object ID.", "type": "string" }, "intervention": { "description": "The intervention configuration for the guard.", "properties": { "action": { "description": "The action to take if conditions are met.", "enum": [ "block", "report", "replace" ], "type": "string" }, "allowedActions": { "description": "The actions this guard is allowed to take.", "items": { "enum": [ "block", "report", "replace" ], "type": "string" }, "maxItems": 10, "type": "array" }, "conditionLogic": { "default": "any", "description": "The action to take if conditions are met.", "enum": [ "any" ], "type": "string" }, "conditions": { "description": "The list of conditions to trigger intervention.", "items": { "description": "The condition to trigger intervention.", "properties": { "comparand": { "anyOf": [ { "type": "boolean" }, { "type": "number" }, { "type": "string" }, { "items": { "description": "The class name to match.", "maxLength": 128, "type": "string" }, "maxItems": 10, "type": "array" } ], "description": "The condition comparand (basis of comparison)." }, "comparator": { "description": "The condition comparator (operator).", "enum": [ "greaterThan", "lessThan", "equals", "notEquals", "is", "isNot", "matches", "doesNotMatch", "contains", "doesNotContain" ], "type": "string" } }, "required": [ "comparand", "comparator" ], "type": "object", "x-versionadded": "v2.35" }, "maxItems": 1, "type": "array" }, "modifyMessage": { "description": "The message to use if the prompt or response is blocked.", "maxLength": 4096, "type": "string" }, "sendNotification": { "description": "The notification event to create if intervention is triggered.", "type": "boolean" } }, "required": [ "action", "conditions", "modifyMessage" ], "type": "object", "x-versionadded": "v2.35" }, "isAgentic": { "default": true, "description": "True if the guard is suitable for agentic workflows only.", "type": "boolean", "x-versionadded": "v2.37" }, "isValid": { "default": true, "description": "True if the guard is fully configured and valid.", "type": "boolean" }, "labels": { "description": "The list of short strings to associate with the template.", "items": { "description": "A short string to associate with the template.", "maxLength": 255, "type": "string" }, "maxItems": 16, "type": "array", "x-versionadded": "v2.37" }, "llmGatewayModelId": { "description": "The LLM Gateway model ID to use as judge.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "llmType": { "description": "The type of LLM used by this guard.", "enum": [ "openAi", "azureOpenAi", "google", "amazon", "datarobot", "nim", "llmGateway" ], "type": [ "string", "null" ] }, "modelInfo": { "description": "The configuration info for guards using deployed models.", "properties": { "classNames": { "description": "The list of class names for multiclass models.", "items": { "description": "The class name.", "maxLength": 128, "type": "string" }, "maxItems": 100, "type": "array" }, "inputColumnName": { "description": "The input column name.", "maxLength": 255, "type": "string" }, "modelId": { "description": "The ID of the registered model for model guards.", "type": [ "string", "null" ] }, "modelName": { "default": "", "description": "The ID of the registered model for .model guards.", "maxLength": 255, "type": "string" }, "outputColumnName": { "description": "The output column name.", "maxLength": 255, "type": "string" }, "replacementTextColumnName": { "default": "", "description": "The name of the output column with replacement text. Required only if intervention.action is `replace`.", "maxLength": 255, "type": "string" }, "targetType": { "description": "The target type.", "enum": [ "Binary", "Regression", "Multiclass", "TextGeneration" ], "type": "string" } }, "required": [ "inputColumnName", "outputColumnName", "targetType" ], "type": "object", "x-versionadded": "v2.35" }, "name": { "description": "The guard template name.", "maxLength": 255, "type": "string" }, "nemoEvaluatorType": { "description": "The guard template \"NeMo Evaluator\" metric type.", "enum": [ "llm_judge", "context_relevance", "response_groundedness", "topic_adherence", "agent_goal_accuracy", "response_relevancy", "faithfulness" ], "type": [ "string", "null" ], "x-versionadded": "v2.41" }, "nemoInfo": { "description": "The configuration info for NeMo guards.", "properties": { "actions": { "default": "", "description": "The NeMo guardrails actions.", "maxLength": 4096, "type": "string" }, "blockedTerms": { "description": "The NeMo guardrails blocked terms list.", "maxLength": 4096, "type": "string" }, "credentialId": { "description": "The NeMo guardrails credential ID.", "type": [ "string", "null" ] }, "llmPrompts": { "default": "", "description": "The NeMo guardrails prompts.", "maxLength": 4096, "type": "string" }, "mainConfig": { "description": "The overall NeMo configuration YAML.", "maxLength": 4096, "type": "string" }, "railsConfig": { "description": "The NeMo guardrails configuration Colang.", "maxLength": 4096, "type": "string" } }, "required": [ "blockedTerms", "mainConfig", "railsConfig" ], "type": "object", "x-versionadded": "v2.35" }, "nemoLlmJudgeConfig": { "description": "The configuration for the LLM Judge metric.", "properties": { "customMetricDirectionality": { "description": "The custom metric directionality for the LLM judge.", "enum": [ "higherIsBetter", "lowerIsBetter" ], "type": "string", "x-versionadded": "v2.42" }, "scoreParsingRegex": { "description": "The regex to parse the score from the LLM judge response.", "maxLength": 1024, "type": "string" }, "systemPrompt": { "description": "The system prompt for the LLM judge.", "maxLength": 8192, "type": "string" }, "userPrompt": { "description": "The user prompt template for the LLM judge.", "maxLength": 8192, "type": "string" } }, "required": [ "customMetricDirectionality", "scoreParsingRegex", "systemPrompt", "userPrompt" ], "type": "object", "x-versionadded": "v2.42" }, "nemoResponseRelevancyConfig": { "description": "The configuration for the Response Relevancy metric.", "properties": { "embeddingDeploymentId": { "description": "The ID of the embedding model deployment.", "type": "string" } }, "required": [ "embeddingDeploymentId" ], "type": "object", "x-versionadded": "v2.42" }, "nemoTopicAdherenceConfig": { "description": "The configuration for the Topic Adherence metric.", "properties": { "metricMode": { "default": "f1", "description": "The metric calculation mode.", "enum": [ "f1", "recall", "precision" ], "type": "string" }, "referenceTopics": { "description": "The list of reference topics.", "items": { "description": "The reference topic.", "maxLength": 512, "type": "string" }, "maxItems": 100, "minItems": 1, "type": "array" } }, "required": [ "referenceTopics" ], "type": "object", "x-versionadded": "v2.42" }, "ootbType": { "description": "The guard template \"Out of the Box\" metric type.", "enum": [ "token_count", "faithfulness", "rouge_1", "agent_goal_accuracy", "agent_goal_accuracy_with_reference", "cost", "task_adherence", "tool_call_accuracy", "agent_guideline_adherence", "agent_latency", "agent_tokens", "agent_cost", "agentGoalAccuracy", "agentGoalAccuracyWithReference", "taskAdherence", "toolCallAccuracy" ], "type": [ "string", "null" ] }, "openaiApiBase": { "description": "The Azure OpenAI API Base URL.", "maxLength": 255, "type": [ "string", "null" ] }, "openaiApiKey": { "description": "Deprecated; use openai_credential instead.", "maxLength": 255, "type": [ "string", "null" ] }, "openaiDeploymentId": { "description": "The OpenAPI deployment ID.", "maxLength": 255, "type": [ "string", "null" ] }, "orgId": { "description": "Organization ID of the user who created the Guard template.", "type": [ "string", "null" ] }, "playgroundOnly": { "description": "Whether the guard is for playground only, or if it can be used in production and playground.", "type": [ "boolean", "null" ], "x-versionadded": "v2.37" }, "productionOnly": { "description": "Whether the guard is for production only, or if it can be used in production and playground.", "type": [ "boolean", "null" ] }, "type": { "description": "The guard template type.", "enum": [ "guardModel", "nemo", "nemoEvaluator", "ootb", "pii", "userModel" ], "type": "string" } }, "required": [ "allowedStages", "createdAt", "description", "id", "name", "type" ], "type": "object", "x-versionadded": "v2.35" }, "maxItems": 200, "type": "array" }, "next": { "description": "The URL of the next page (if null, there is no next page).", "format": "uri", "type": [ "string", "null" ] }, "previous": { "description": "The URL of the previous page (if null, there is no previous page).", "format": "uri", "type": [ "string", "null" ] }, "totalCount": { "description": "The total number of items across all pages.", "type": "integer" } }, "required": [ "data", "next", "previous", "totalCount" ], "type": "object", "x-versionadded": "v2.35" } ``` ### Responses | Status | Meaning | Description | Schema | | --- | --- | --- | --- | | 200 | OK | none | GuardTemplateListResponse | | 400 | Bad Request | Request invalid, refer to messages for detail. | None | | 404 | Not Found | Missing feature flag. | None | ## Retrieve information about a guard template by template ID Operation path: `GET /api/v2/guardTemplates/{templateId}/` Authentication requirements: `BearerAuth` Retrieve information about a guard template. ### Parameters | Name | In | Type | Required | Description | | --- | --- | --- | --- | --- | | templateId | path | string | true | The ID of the template. | ### Example responses > 200 Response ``` { "properties": { "allowedActions": { "description": "The actions this guard is allowed to take.", "items": { "enum": [ "block", "report", "replace" ], "type": "string" }, "maxItems": 10, "type": "array", "x-versionadded": "v2.36" }, "allowedStages": { "description": "The stages where the guard can run.", "items": { "enum": [ "prompt", "response" ], "type": "string" }, "maxItems": 16, "type": "array" }, "awsModel": { "description": "AWS model.", "enum": [ "amazon-titan", "anthropic-claude-2", "anthropic-claude-3-haiku", "anthropic-claude-3-sonnet", "anthropic-claude-3-opus", "anthropic-claude-3.5-sonnet-v1", "anthropic-claude-3.5-sonnet-v2", "amazon-nova-lite", "amazon-nova-micro", "amazon-nova-pro" ], "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "awsRegion": { "description": "AWS model region.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "createdAt": { "description": "When the template was created.", "format": "date-time", "type": "string" }, "creatorId": { "description": "The ID of the user who created the guard template.", "type": [ "string", "null" ] }, "creatorName": { "description": "The name of the user who created the guard template.", "maxLength": 1000, "type": "string" }, "description": { "description": "The guard template description.", "maxLength": 4096, "type": "string" }, "errorMessage": { "description": "Error message if the guard configuration is invalid.", "type": [ "string", "null" ] }, "googleModel": { "description": "Google model.", "enum": [ "chat-bison", "google-gemini-1.5-flash", "google-gemini-1.5-pro" ], "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "googleRegion": { "description": "Google model region.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "id": { "description": "The guard template object ID.", "type": "string" }, "intervention": { "description": "The intervention configuration for the guard.", "properties": { "action": { "description": "The action to take if conditions are met.", "enum": [ "block", "report", "replace" ], "type": "string" }, "allowedActions": { "description": "The actions this guard is allowed to take.", "items": { "enum": [ "block", "report", "replace" ], "type": "string" }, "maxItems": 10, "type": "array" }, "conditionLogic": { "default": "any", "description": "The action to take if conditions are met.", "enum": [ "any" ], "type": "string" }, "conditions": { "description": "The list of conditions to trigger intervention.", "items": { "description": "The condition to trigger intervention.", "properties": { "comparand": { "anyOf": [ { "type": "boolean" }, { "type": "number" }, { "type": "string" }, { "items": { "description": "The class name to match.", "maxLength": 128, "type": "string" }, "maxItems": 10, "type": "array" } ], "description": "The condition comparand (basis of comparison)." }, "comparator": { "description": "The condition comparator (operator).", "enum": [ "greaterThan", "lessThan", "equals", "notEquals", "is", "isNot", "matches", "doesNotMatch", "contains", "doesNotContain" ], "type": "string" } }, "required": [ "comparand", "comparator" ], "type": "object", "x-versionadded": "v2.35" }, "maxItems": 1, "type": "array" }, "modifyMessage": { "description": "The message to use if the prompt or response is blocked.", "maxLength": 4096, "type": "string" }, "sendNotification": { "description": "The notification event to create if intervention is triggered.", "type": "boolean" } }, "required": [ "action", "conditions", "modifyMessage" ], "type": "object", "x-versionadded": "v2.35" }, "isAgentic": { "default": true, "description": "True if the guard is suitable for agentic workflows only.", "type": "boolean", "x-versionadded": "v2.37" }, "isValid": { "default": true, "description": "True if the guard is fully configured and valid.", "type": "boolean" }, "labels": { "description": "The list of short strings to associate with the template.", "items": { "description": "A short string to associate with the template.", "maxLength": 255, "type": "string" }, "maxItems": 16, "type": "array", "x-versionadded": "v2.37" }, "llmGatewayModelId": { "description": "The LLM Gateway model ID to use as judge.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "llmType": { "description": "The type of LLM used by this guard.", "enum": [ "openAi", "azureOpenAi", "google", "amazon", "datarobot", "nim", "llmGateway" ], "type": [ "string", "null" ] }, "modelInfo": { "description": "The configuration info for guards using deployed models.", "properties": { "classNames": { "description": "The list of class names for multiclass models.", "items": { "description": "The class name.", "maxLength": 128, "type": "string" }, "maxItems": 100, "type": "array" }, "inputColumnName": { "description": "The input column name.", "maxLength": 255, "type": "string" }, "modelId": { "description": "The ID of the registered model for model guards.", "type": [ "string", "null" ] }, "modelName": { "default": "", "description": "The ID of the registered model for .model guards.", "maxLength": 255, "type": "string" }, "outputColumnName": { "description": "The output column name.", "maxLength": 255, "type": "string" }, "replacementTextColumnName": { "default": "", "description": "The name of the output column with replacement text. Required only if intervention.action is `replace`.", "maxLength": 255, "type": "string" }, "targetType": { "description": "The target type.", "enum": [ "Binary", "Regression", "Multiclass", "TextGeneration" ], "type": "string" } }, "required": [ "inputColumnName", "outputColumnName", "targetType" ], "type": "object", "x-versionadded": "v2.35" }, "name": { "description": "The guard template name.", "maxLength": 255, "type": "string" }, "nemoEvaluatorType": { "description": "The guard template \"NeMo Evaluator\" metric type.", "enum": [ "llm_judge", "context_relevance", "response_groundedness", "topic_adherence", "agent_goal_accuracy", "response_relevancy", "faithfulness" ], "type": [ "string", "null" ], "x-versionadded": "v2.41" }, "nemoInfo": { "description": "The configuration info for NeMo guards.", "properties": { "actions": { "default": "", "description": "The NeMo guardrails actions.", "maxLength": 4096, "type": "string" }, "blockedTerms": { "description": "The NeMo guardrails blocked terms list.", "maxLength": 4096, "type": "string" }, "credentialId": { "description": "The NeMo guardrails credential ID.", "type": [ "string", "null" ] }, "llmPrompts": { "default": "", "description": "The NeMo guardrails prompts.", "maxLength": 4096, "type": "string" }, "mainConfig": { "description": "The overall NeMo configuration YAML.", "maxLength": 4096, "type": "string" }, "railsConfig": { "description": "The NeMo guardrails configuration Colang.", "maxLength": 4096, "type": "string" } }, "required": [ "blockedTerms", "mainConfig", "railsConfig" ], "type": "object", "x-versionadded": "v2.35" }, "nemoLlmJudgeConfig": { "description": "The configuration for the LLM Judge metric.", "properties": { "customMetricDirectionality": { "description": "The custom metric directionality for the LLM judge.", "enum": [ "higherIsBetter", "lowerIsBetter" ], "type": "string", "x-versionadded": "v2.42" }, "scoreParsingRegex": { "description": "The regex to parse the score from the LLM judge response.", "maxLength": 1024, "type": "string" }, "systemPrompt": { "description": "The system prompt for the LLM judge.", "maxLength": 8192, "type": "string" }, "userPrompt": { "description": "The user prompt template for the LLM judge.", "maxLength": 8192, "type": "string" } }, "required": [ "customMetricDirectionality", "scoreParsingRegex", "systemPrompt", "userPrompt" ], "type": "object", "x-versionadded": "v2.42" }, "nemoResponseRelevancyConfig": { "description": "The configuration for the Response Relevancy metric.", "properties": { "embeddingDeploymentId": { "description": "The ID of the embedding model deployment.", "type": "string" } }, "required": [ "embeddingDeploymentId" ], "type": "object", "x-versionadded": "v2.42" }, "nemoTopicAdherenceConfig": { "description": "The configuration for the Topic Adherence metric.", "properties": { "metricMode": { "default": "f1", "description": "The metric calculation mode.", "enum": [ "f1", "recall", "precision" ], "type": "string" }, "referenceTopics": { "description": "The list of reference topics.", "items": { "description": "The reference topic.", "maxLength": 512, "type": "string" }, "maxItems": 100, "minItems": 1, "type": "array" } }, "required": [ "referenceTopics" ], "type": "object", "x-versionadded": "v2.42" }, "ootbType": { "description": "The guard template \"Out of the Box\" metric type.", "enum": [ "token_count", "faithfulness", "rouge_1", "agent_goal_accuracy", "agent_goal_accuracy_with_reference", "cost", "task_adherence", "tool_call_accuracy", "agent_guideline_adherence", "agent_latency", "agent_tokens", "agent_cost", "agentGoalAccuracy", "agentGoalAccuracyWithReference", "taskAdherence", "toolCallAccuracy" ], "type": [ "string", "null" ] }, "openaiApiBase": { "description": "The Azure OpenAI API Base URL.", "maxLength": 255, "type": [ "string", "null" ] }, "openaiApiKey": { "description": "Deprecated; use openai_credential instead.", "maxLength": 255, "type": [ "string", "null" ] }, "openaiDeploymentId": { "description": "The OpenAPI deployment ID.", "maxLength": 255, "type": [ "string", "null" ] }, "orgId": { "description": "Organization ID of the user who created the Guard template.", "type": [ "string", "null" ] }, "playgroundOnly": { "description": "Whether the guard is for playground only, or if it can be used in production and playground.", "type": [ "boolean", "null" ], "x-versionadded": "v2.37" }, "productionOnly": { "description": "Whether the guard is for production only, or if it can be used in production and playground.", "type": [ "boolean", "null" ] }, "type": { "description": "The guard template type.", "enum": [ "guardModel", "nemo", "nemoEvaluator", "ootb", "pii", "userModel" ], "type": "string" } }, "required": [ "allowedStages", "createdAt", "description", "id", "name", "type" ], "type": "object", "x-versionadded": "v2.35" } ``` ### Responses | Status | Meaning | Description | Schema | | --- | --- | --- | --- | | 200 | OK | none | GuardTemplateRetrieveResponse | | 404 | Not Found | Either the template does not exist or the required feature flag is missing. | None | ## The list of supported LLMs Operation path: `GET /api/v2/moderationSupportedLlms/` Authentication requirements: `BearerAuth` The list of supported LLMs for moderation. ### Parameters | Name | In | Type | Required | Description | | --- | --- | --- | --- | --- | | offset | query | integer | false | This many results will be skipped. | | limit | query | integer | false | At most this many results are returned. | ### Example responses > 200 Response ``` { "properties": { "count": { "description": "The number of items returned on this page.", "type": "integer" }, "data": { "description": "The list of supported LLMs for moderation.", "items": { "properties": { "description": { "description": "The description of this LLM.", "maxLength": 1024, "type": "string" }, "id": { "description": "The ID for this LLM.", "maxLength": 1024, "type": "string" }, "llmType": { "description": "The general category of this LLM.", "maxLength": 1024, "type": "string" }, "model": { "description": "The specific model of this LLM.", "maxLength": 1024, "type": "string" }, "name": { "description": "The display name of this LLM.", "maxLength": 1024, "type": "string" }, "provider": { "description": "The provider of access to this LLM.", "maxLength": 1024, "type": "string" }, "vendor": { "description": "The vendor of this LLM.", "maxLength": 1024, "type": "string" } }, "required": [ "description", "id", "llmType", "model", "name", "provider", "vendor" ], "type": "object", "x-versionadded": "v2.35" }, "maxItems": 200, "type": "array" }, "next": { "description": "The URL of the next page (if null, there is no next page).", "format": "uri", "type": [ "string", "null" ] }, "previous": { "description": "The URL of the previous page (if null, there is no previous page).", "format": "uri", "type": [ "string", "null" ] }, "totalCount": { "description": "The total number of items across all pages.", "type": "integer" } }, "required": [ "data", "next", "previous", "totalCount" ], "type": "object", "x-versionadded": "v2.35" } ``` ### Responses | Status | Meaning | Description | Schema | | --- | --- | --- | --- | | 200 | OK | none | SupportedLlmListResponse | | 400 | Bad Request | Request invalid, refer to messages for detail. | None | | 404 | Not Found | Missing feature flag. | None | ## Get the overall moderation configuration Operation path: `GET /api/v2/overallModerationConfiguration/` Authentication requirements: `BearerAuth` Get the overall moderation configuration for an entity. ### Parameters | Name | In | Type | Required | Description | | --- | --- | --- | --- | --- | | entityId | query | string | true | Retrieve the overall moderation configuration for the given entity ID. | | entityType | query | string | true | The entity type of the given entity ID. | ### Enumerated Values | Parameter | Value | | --- | --- | | entityType | [customModel, customModelVersion, playground] | ### Example responses > 200 Response ``` { "properties": { "entityId": { "description": "The ID of the custom model or playground for this configuration.", "type": "string" }, "entityType": { "description": "The type of the associated entity.", "enum": [ "customModel", "customModelVersion", "playground" ], "type": "string" }, "nemoEvaluatorDeploymentId": { "description": "The ID of the NeMo Evaluator deployment to use for all NeMo Evaluator guards.", "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "timeoutAction": { "description": "The action to take if timeout occurs.", "enum": [ "block", "score" ], "type": "string" }, "timeoutSec": { "description": "The timeout value in seconds for any guard.", "minimum": 2, "type": "integer" }, "updatedAt": { "description": "When the configuration was updated.", "format": "date-time", "type": "string" }, "updaterId": { "description": "The ID of the user who updated the configuration.", "type": [ "string", "null" ] } }, "required": [ "entityId", "entityType", "timeoutAction", "timeoutSec", "updaterId" ], "type": "object", "x-versionadded": "v2.35" } ``` ### Responses | Status | Meaning | Description | Schema | | --- | --- | --- | --- | | 200 | OK | none | OverallModerationConfigurationResponse | | 404 | Not Found | User permissions problem. | None | ## Update the overall moderation configuration Operation path: `PATCH /api/v2/overallModerationConfiguration/` Authentication requirements: `BearerAuth` Update the overall moderation configuration for an entity. ### Body parameter ``` { "properties": { "entityId": { "description": "The ID of the custom model or playground for this configuration.", "type": "string" }, "entityType": { "description": "The type of the associated entity.", "enum": [ "customModel", "customModelVersion", "playground" ], "type": "string" }, "nemoEvaluatorDeploymentId": { "description": "The ID of the NeMo Evaluator deployment to use for all NeMo Evaluator guards.", "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "timeoutAction": { "description": "The action to take if timeout occurs.", "enum": [ "block", "score" ], "type": "string" }, "timeoutSec": { "description": "The timeout value in seconds for any guard.", "minimum": 0, "type": "integer" } }, "required": [ "entityId", "entityType", "timeoutAction", "timeoutSec" ], "type": "object", "x-versionadded": "v2.35" } ``` ### Parameters | Name | In | Type | Required | Description | | --- | --- | --- | --- | --- | | entityId | path | string | true | Retrieve the overall moderation configuration for the given entity ID. | | entityType | path | string | true | The entity type of the given entity ID. | | body | body | OverallModerationConfigurationUpdate | false | none | ### Enumerated Values | Parameter | Value | | --- | --- | | entityType | [customModel, customModelVersion, playground] | ### Example responses > 200 Response ``` { "properties": { "entityId": { "description": "The ID of the custom model or playground for this configuration.", "type": "string" }, "entityType": { "description": "The type of the associated entity.", "enum": [ "customModel", "customModelVersion", "playground" ], "type": "string" }, "nemoEvaluatorDeploymentId": { "description": "The ID of the NeMo Evaluator deployment to use for all NeMo Evaluator guards.", "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "timeoutAction": { "description": "The action to take if timeout occurs.", "enum": [ "block", "score" ], "type": "string" }, "timeoutSec": { "description": "The timeout value in seconds for any guard.", "minimum": 2, "type": "integer" }, "updatedAt": { "description": "When the configuration was updated.", "format": "date-time", "type": "string" }, "updaterId": { "description": "The ID of the user who updated the configuration.", "type": [ "string", "null" ] } }, "required": [ "entityId", "entityType", "timeoutAction", "timeoutSec", "updaterId" ], "type": "object", "x-versionadded": "v2.35" } ``` ### Responses | Status | Meaning | Description | Schema | | --- | --- | --- | --- | | 200 | OK | none | OverallModerationConfigurationResponse | | 404 | Not Found | Either the resource does not exist or the user does not have permission to create the configuration. | None | # Schemas ## DeploymentAndGuardResponse ``` { "properties": { "configurationId": { "description": "ID of guard configuration.", "type": "string" }, "deploymentId": { "description": "ID of guard model deployment.", "type": "string" }, "name": { "description": "Name of guard configuration.", "type": "string" } }, "required": [ "configurationId", "deploymentId", "name" ], "type": "object", "x-versionadded": "v2.35" } ``` ### Properties | Name | Type | Required | Restrictions | Description | | --- | --- | --- | --- | --- | | configurationId | string | true | | ID of guard configuration. | | deploymentId | string | true | | ID of guard model deployment. | | name | string | true | | Name of guard configuration. | ## GuardAdditionalConfig ``` { "description": "Additional configuration for the guard.", "properties": { "agentGuideline": { "description": "How to calculate agent guideline adherence for this guard.", "maxLength": 4096, "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "cost": { "description": "How to calculate cost information for this guard.", "properties": { "currency": { "description": "ISO 4217 Currency code for display.", "maxLength": 255, "type": "string" }, "inputPrice": { "description": "Cost per unit measure of tokens for input.", "minimum": 0, "type": "number" }, "inputUnit": { "description": "Number of tokens related to input price.", "minimum": 0, "type": "integer" }, "outputPrice": { "description": "Cost per unit measure of tokens for output.", "minimum": 0, "type": "number" }, "outputUnit": { "description": "Number of tokens related to output price.", "minimum": 0, "type": "integer" } }, "required": [ "currency", "inputPrice", "inputUnit", "outputPrice", "outputUnit" ], "type": "object", "x-versionadded": "v2.37" }, "toolCall": { "description": "How to calculate tool call metrics for this guard.", "properties": { "comparisonMetric": { "description": "How tool calls should be compared for metrics calculations.", "enum": [ "exactMatch", "doNotCompare" ], "type": "string" } }, "required": [ "comparisonMetric" ], "type": "object", "x-versionadded": "v2.37" } }, "type": "object", "x-versionadded": "v2.37" } ``` Additional configuration for the guard. ### Properties | Name | Type | Required | Restrictions | Description | | --- | --- | --- | --- | --- | | agentGuideline | string,null | false | maxLength: 4096 | How to calculate agent guideline adherence for this guard. | | cost | GuardConfigurationCostInfo | false | | How to calculate cost information for this guard. | | toolCall | GuardToolCallInfo | false | | How to calculate tool call metrics for this guard. | ## GuardConditionResponse ``` { "description": "The condition to trigger intervention.", "properties": { "comparand": { "anyOf": [ { "type": "boolean" }, { "type": "number" }, { "type": "string" }, { "items": { "description": "The class name to match.", "maxLength": 128, "type": "string" }, "maxItems": 10, "type": "array" } ], "description": "The condition comparand (basis of comparison)." }, "comparator": { "description": "The condition comparator (operator).", "enum": [ "greaterThan", "lessThan", "equals", "notEquals", "is", "isNot", "matches", "doesNotMatch", "contains", "doesNotContain" ], "type": "string" } }, "required": [ "comparand", "comparator" ], "type": "object", "x-versionadded": "v2.35" } ``` The condition to trigger intervention. ### Properties | Name | Type | Required | Restrictions | Description | | --- | --- | --- | --- | --- | | comparand | any | true | | The condition comparand (basis of comparison). | anyOf | Name | Type | Required | Restrictions | Description | | --- | --- | --- | --- | --- | | » anonymous | boolean | false | | none | or | Name | Type | Required | Restrictions | Description | | --- | --- | --- | --- | --- | | » anonymous | number | false | | none | or | Name | Type | Required | Restrictions | Description | | --- | --- | --- | --- | --- | | » anonymous | string | false | | none | or | Name | Type | Required | Restrictions | Description | | --- | --- | --- | --- | --- | | » anonymous | [string] | false | maxItems: 10 | none | continued | Name | Type | Required | Restrictions | Description | | --- | --- | --- | --- | --- | | comparator | string | true | | The condition comparator (operator). | ### Enumerated Values | Property | Value | | --- | --- | | comparator | [greaterThan, lessThan, equals, notEquals, is, isNot, matches, doesNotMatch, contains, doesNotContain] | ## GuardConfigurationConditionResponse ``` { "description": "The condition to trigger intervention.", "properties": { "comparand": { "anyOf": [ { "type": "boolean" }, { "type": "number" }, { "type": "string" }, { "items": { "type": "string" }, "maxItems": 10, "type": "array" } ], "description": "The condition comparand (basis of comparison)." }, "comparator": { "description": "The condition comparator (operator).", "enum": [ "greaterThan", "lessThan", "equals", "notEquals", "is", "isNot", "matches", "doesNotMatch", "contains", "doesNotContain" ], "type": "string" } }, "required": [ "comparand", "comparator" ], "type": "object", "x-versionadded": "v2.35" } ``` The condition to trigger intervention. ### Properties | Name | Type | Required | Restrictions | Description | | --- | --- | --- | --- | --- | | comparand | any | true | | The condition comparand (basis of comparison). | anyOf | Name | Type | Required | Restrictions | Description | | --- | --- | --- | --- | --- | | » anonymous | boolean | false | | none | or | Name | Type | Required | Restrictions | Description | | --- | --- | --- | --- | --- | | » anonymous | number | false | | none | or | Name | Type | Required | Restrictions | Description | | --- | --- | --- | --- | --- | | » anonymous | string | false | | none | or | Name | Type | Required | Restrictions | Description | | --- | --- | --- | --- | --- | | » anonymous | [string] | false | maxItems: 10 | none | continued | Name | Type | Required | Restrictions | Description | | --- | --- | --- | --- | --- | | comparator | string | true | | The condition comparator (operator). | ### Enumerated Values | Property | Value | | --- | --- | | comparator | [greaterThan, lessThan, equals, notEquals, is, isNot, matches, doesNotMatch, contains, doesNotContain] | ## GuardConfigurationCostInfo ``` { "description": "How to calculate cost information for this guard.", "properties": { "currency": { "description": "ISO 4217 Currency code for display.", "maxLength": 255, "type": "string" }, "inputPrice": { "description": "Cost per unit measure of tokens for input.", "minimum": 0, "type": "number" }, "inputUnit": { "description": "Number of tokens related to input price.", "minimum": 0, "type": "integer" }, "outputPrice": { "description": "Cost per unit measure of tokens for output.", "minimum": 0, "type": "number" }, "outputUnit": { "description": "Number of tokens related to output price.", "minimum": 0, "type": "integer" } }, "required": [ "currency", "inputPrice", "inputUnit", "outputPrice", "outputUnit" ], "type": "object", "x-versionadded": "v2.37" } ``` How to calculate cost information for this guard. ### Properties | Name | Type | Required | Restrictions | Description | | --- | --- | --- | --- | --- | | currency | string | true | maxLength: 255 | ISO 4217 Currency code for display. | | inputPrice | number | true | minimum: 0 | Cost per unit measure of tokens for input. | | inputUnit | integer | true | minimum: 0 | Number of tokens related to input price. | | outputPrice | number | true | minimum: 0 | Cost per unit measure of tokens for output. | | outputUnit | integer | true | minimum: 0 | Number of tokens related to output price. | ## GuardConfigurationCreate ``` { "properties": { "additionalGuardConfig": { "description": "Additional configuration for the guard.", "properties": { "agentGuideline": { "description": "How to calculate agent guideline adherence for this guard.", "maxLength": 4096, "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "cost": { "description": "How to calculate cost information for this guard.", "properties": { "currency": { "description": "ISO 4217 Currency code for display.", "maxLength": 255, "type": "string" }, "inputPrice": { "description": "Cost per unit measure of tokens for input.", "minimum": 0, "type": "number" }, "inputUnit": { "description": "Number of tokens related to input price.", "minimum": 0, "type": "integer" }, "outputPrice": { "description": "Cost per unit measure of tokens for output.", "minimum": 0, "type": "number" }, "outputUnit": { "description": "Number of tokens related to output price.", "minimum": 0, "type": "integer" } }, "required": [ "currency", "inputPrice", "inputUnit", "outputPrice", "outputUnit" ], "type": "object", "x-versionadded": "v2.37" }, "toolCall": { "description": "How to calculate tool call metrics for this guard.", "properties": { "comparisonMetric": { "description": "How tool calls should be compared for metrics calculations.", "enum": [ "exactMatch", "doNotCompare" ], "type": "string" } }, "required": [ "comparisonMetric" ], "type": "object", "x-versionadded": "v2.37" } }, "type": "object", "x-versionadded": "v2.37" }, "allowedActions": { "description": "The actions this guard is allowed to take.", "items": { "enum": [ "block", "report", "replace" ], "type": "string" }, "maxItems": 10, "type": "array", "x-versionadded": "v2.36" }, "awsAccount": { "description": "The ID of the user credential containing an AWS account.", "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "awsModel": { "description": "AWS model.", "enum": [ "amazon-titan", "anthropic-claude-2", "anthropic-claude-3-haiku", "anthropic-claude-3-sonnet", "anthropic-claude-3-opus", "anthropic-claude-3.5-sonnet-v1", "anthropic-claude-3.5-sonnet-v2", "amazon-nova-lite", "amazon-nova-micro", "amazon-nova-pro" ], "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "awsRegion": { "description": "AWS model region.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "deploymentId": { "description": "The ID of the deployed model for model guards.", "type": [ "string", "null" ] }, "description": { "description": "Guard configuration description", "maxLength": 4096, "type": "string" }, "entityId": { "description": "The ID of the custom model or playground for this guard.", "type": "string" }, "entityType": { "description": "The type of the associated entity.", "enum": [ "customModel", "customModelVersion", "playground" ], "type": "string" }, "googleModel": { "description": "Google model.", "enum": [ "chat-bison", "google-gemini-1.5-flash", "google-gemini-1.5-pro" ], "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "googleRegion": { "description": "Google model region.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "googleServiceAccount": { "description": "The ID of the user credential containing a Google service account.", "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "intervention": { "description": "Intervention configuration for the guard.", "properties": { "action": { "description": "The action to take if conditions are met.", "enum": [ "block", "report", "replace" ], "type": "string" }, "allowedActions": { "description": "The actions this guard is allowed to take.", "items": { "enum": [ "block", "report", "replace" ], "type": "string" }, "maxItems": 10, "type": "array" }, "conditionLogic": { "default": "any", "description": "The action to take if conditions are met.", "enum": [ "any" ], "type": "string" }, "conditions": { "description": "The list of conditions to trigger intervention.", "items": { "description": "The condition to trigger intervention.", "properties": { "comparand": { "anyOf": [ { "type": "boolean" }, { "type": "number" }, { "type": "string" }, { "items": { "type": "string" }, "maxItems": 10, "type": "array" } ], "description": "The condition comparand (basis of comparison)." }, "comparator": { "description": "The condition comparator (operator).", "enum": [ "greaterThan", "lessThan", "equals", "notEquals", "is", "isNot", "matches", "doesNotMatch", "contains", "doesNotContain" ], "type": "string" } }, "required": [ "comparand", "comparator" ], "type": "object", "x-versionadded": "v2.35" }, "maxItems": 1, "type": "array" }, "message": { "description": "The message to use if the prompt or response is blocked.", "maxLength": 4096, "type": "string" }, "sendNotification": { "default": false, "description": "The notification event to create if intervention is triggered.", "type": "boolean" } }, "required": [ "action", "conditions", "message" ], "type": "object", "x-versionadded": "v2.35" }, "llmGatewayModelId": { "description": "The LLM Gateway model ID to use as judge.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "llmType": { "description": "The type of LLM used by this guard.", "enum": [ "openAi", "azureOpenAi", "google", "amazon", "datarobot", "nim", "llmGateway" ], "type": [ "string", "null" ] }, "modelInfo": { "description": "The configuration info for guards using deployed models.", "properties": { "classNames": { "description": "The list of class names for multiclass models.", "items": { "description": "The class name.", "maxLength": 128, "type": "string" }, "maxItems": 100, "type": "array" }, "inputColumnName": { "description": "The input column name.", "maxLength": 255, "type": "string" }, "modelId": { "description": "The ID of the registered model for model guards.", "type": [ "string", "null" ] }, "modelName": { "default": "", "description": "The ID of the registered model for .model guards.", "maxLength": 255, "type": "string" }, "outputColumnName": { "description": "The output column name.", "maxLength": 255, "type": "string" }, "replacementTextColumnName": { "default": "", "description": "The name of the output column with replacement text. Required only if intervention.action is `replace`.", "maxLength": 255, "type": "string" }, "targetType": { "description": "The target type.", "enum": [ "Binary", "Regression", "Multiclass", "TextGeneration" ], "type": [ "string", "null" ] } }, "required": [ "inputColumnName", "outputColumnName" ], "type": "object", "x-versionadded": "v2.35" }, "name": { "description": "Guard configuration name", "maxLength": 255, "type": "string" }, "nemoInfo": { "description": "The configuration info for NeMo guards.", "properties": { "actions": { "description": "The NeMo guardrails actions file.", "maxLength": 4096, "type": "string" }, "blockedTerms": { "description": "The NeMo guardrails blocked terms list.", "maxLength": 4096, "type": "string" }, "credentialId": { "description": "The NeMo guardrails credential ID (deprecated; use \"openai_credential\").", "type": [ "string", "null" ] }, "llmPrompts": { "description": "The NeMo guardrails prompts.", "maxLength": 4096, "type": "string" }, "mainConfig": { "description": "The overall NeMo configuration YAML.", "maxLength": 4096, "type": "string" }, "railsConfig": { "description": "The NeMo guardrails configuration Colang.", "maxLength": 4096, "type": "string" } }, "required": [ "blockedTerms", "mainConfig", "railsConfig" ], "type": "object", "x-versionadded": "v2.35" }, "nemoLlmJudgeConfig": { "description": "The configuration for the LLM Judge metric.", "properties": { "customMetricDirectionality": { "description": "The custom metric directionality for the LLM judge.", "enum": [ "higherIsBetter", "lowerIsBetter" ], "type": "string", "x-versionadded": "v2.42" }, "scoreParsingRegex": { "description": "The regex to parse the score from the LLM judge response.", "maxLength": 1024, "type": "string" }, "systemPrompt": { "description": "The system prompt for the LLM judge.", "maxLength": 8192, "type": "string" }, "userPrompt": { "description": "The user prompt template for the LLM judge.", "maxLength": 8192, "type": "string" } }, "required": [ "customMetricDirectionality", "scoreParsingRegex", "systemPrompt", "userPrompt" ], "type": "object", "x-versionadded": "v2.42" }, "nemoResponseRelevancyConfig": { "description": "The configuration for the Response Relevancy metric.", "properties": { "embeddingDeploymentId": { "description": "The ID of the embedding model deployment.", "type": "string" } }, "required": [ "embeddingDeploymentId" ], "type": "object", "x-versionadded": "v2.42" }, "nemoTopicAdherenceConfig": { "description": "The configuration for the Topic Adherence metric.", "properties": { "metricMode": { "default": "f1", "description": "The metric calculation mode.", "enum": [ "f1", "recall", "precision" ], "type": "string" }, "referenceTopics": { "description": "The list of reference topics.", "items": { "description": "The reference topic.", "maxLength": 512, "type": "string" }, "maxItems": 100, "minItems": 1, "type": "array" } }, "required": [ "referenceTopics" ], "type": "object", "x-versionadded": "v2.42" }, "openaiApiBase": { "description": "The Azure OpenAI API Base URL.", "maxLength": 255, "type": [ "string", "null" ] }, "openaiApiKey": { "description": "Deprecated; use openai_credential instead.", "maxLength": 255, "type": [ "string", "null" ] }, "openaiCredential": { "description": "The ID of the user credential containing an OpenAI token.", "type": [ "string", "null" ] }, "openaiDeploymentId": { "description": "The OpenAPI deployment ID.", "maxLength": 255, "type": [ "string", "null" ] }, "stages": { "description": "The stages where the guard can run.", "items": { "enum": [ "prompt", "response" ], "type": "string" }, "maxItems": 16, "type": "array" }, "templateId": { "description": "The ID of the template this guard is based on.", "type": "string" } }, "required": [ "entityId", "entityType", "name", "stages", "templateId" ], "type": "object", "x-versionadded": "v2.35" } ``` ### Properties | Name | Type | Required | Restrictions | Description | | --- | --- | --- | --- | --- | | additionalGuardConfig | GuardAdditionalConfig | false | | Additional configuration for the guard. | | allowedActions | [string] | false | maxItems: 10 | The actions this guard is allowed to take. | | awsAccount | string,null | false | | The ID of the user credential containing an AWS account. | | awsModel | string,null | false | | AWS model. | | awsRegion | string,null | false | maxLength: 255 | AWS model region. | | deploymentId | string,null | false | | The ID of the deployed model for model guards. | | description | string | false | maxLength: 4096 | Guard configuration description | | entityId | string | true | | The ID of the custom model or playground for this guard. | | entityType | string | true | | The type of the associated entity. | | googleModel | string,null | false | | Google model. | | googleRegion | string,null | false | maxLength: 255 | Google model region. | | googleServiceAccount | string,null | false | | The ID of the user credential containing a Google service account. | | intervention | GuardConfigurationInterventionResponse | false | | Intervention configuration for the guard. | | llmGatewayModelId | string,null | false | maxLength: 255 | The LLM Gateway model ID to use as judge. | | llmType | string,null | false | | The type of LLM used by this guard. | | modelInfo | GuardConfigurationPayloadModelInfo | false | | The configuration info for guards using deployed models. | | name | string | true | maxLength: 255 | Guard configuration name | | nemoInfo | GuardConfigurationNemoInfoResponse | false | | The configuration info for NeMo guards. | | nemoLlmJudgeConfig | GuardNemoLlmJudgeConfigResponse | false | | The configuration for the LLM Judge metric. | | nemoResponseRelevancyConfig | GuardNemoResponseRelevancyConfigResponse | false | | The configuration for the Response Relevancy metric. | | nemoTopicAdherenceConfig | GuardNemoTopicAdherenceConfigResponse | false | | The configuration for the Topic Adherence metric. | | openaiApiBase | string,null | false | maxLength: 255 | The Azure OpenAI API Base URL. | | openaiApiKey | string,null | false | maxLength: 255 | Deprecated; use openai_credential instead. | | openaiCredential | string,null | false | | The ID of the user credential containing an OpenAI token. | | openaiDeploymentId | string,null | false | maxLength: 255 | The OpenAPI deployment ID. | | stages | [string] | true | maxItems: 16 | The stages where the guard can run. | | templateId | string | true | | The ID of the template this guard is based on. | ### Enumerated Values | Property | Value | | --- | --- | | awsModel | [amazon-titan, anthropic-claude-2, anthropic-claude-3-haiku, anthropic-claude-3-sonnet, anthropic-claude-3-opus, anthropic-claude-3.5-sonnet-v1, anthropic-claude-3.5-sonnet-v2, amazon-nova-lite, amazon-nova-micro, amazon-nova-pro] | | entityType | [customModel, customModelVersion, playground] | | googleModel | [chat-bison, google-gemini-1.5-flash, google-gemini-1.5-pro] | | llmType | [openAi, azureOpenAi, google, amazon, datarobot, nim, llmGateway] | ## GuardConfigurationFullPost ``` { "description": "Complete guard configuration to push", "properties": { "additionalGuardConfig": { "description": "Additional configuration for the guard.", "properties": { "agentGuideline": { "description": "How to calculate agent guideline adherence for this guard.", "maxLength": 4096, "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "cost": { "description": "How to calculate cost information for this guard.", "properties": { "currency": { "description": "ISO 4217 Currency code for display.", "maxLength": 255, "type": "string" }, "inputPrice": { "description": "Cost per unit measure of tokens for input.", "minimum": 0, "type": "number" }, "inputUnit": { "description": "Number of tokens related to input price.", "minimum": 0, "type": "integer" }, "outputPrice": { "description": "Cost per unit measure of tokens for output.", "minimum": 0, "type": "number" }, "outputUnit": { "description": "Number of tokens related to output price.", "minimum": 0, "type": "integer" } }, "required": [ "currency", "inputPrice", "inputUnit", "outputPrice", "outputUnit" ], "type": "object", "x-versionadded": "v2.37" }, "toolCall": { "description": "How to calculate tool call metrics for this guard.", "properties": { "comparisonMetric": { "description": "How tool calls should be compared for metrics calculations.", "enum": [ "exactMatch", "doNotCompare" ], "type": "string" } }, "required": [ "comparisonMetric" ], "type": "object", "x-versionadded": "v2.37" } }, "type": "object", "x-versionadded": "v2.37" }, "allowedActions": { "description": "The actions this guard is allowed to take.", "items": { "enum": [ "block", "report", "replace" ], "type": "string" }, "maxItems": 10, "type": "array", "x-versionadded": "v2.36" }, "awsAccount": { "description": "The ID of the user credential containing an AWS account.", "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "awsModel": { "description": "AWS model.", "enum": [ "amazon-titan", "anthropic-claude-2", "anthropic-claude-3-haiku", "anthropic-claude-3-sonnet", "anthropic-claude-3-opus", "anthropic-claude-3.5-sonnet-v1", "anthropic-claude-3.5-sonnet-v2", "amazon-nova-lite", "amazon-nova-micro", "amazon-nova-pro" ], "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "awsRegion": { "description": "AWS model region.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "deploymentId": { "description": "The ID of the deployed model for model guards.", "type": [ "string", "null" ] }, "description": { "description": "Guard configuration description", "maxLength": 4096, "type": "string" }, "errorMessage": { "description": "Error message if the guard configuration is invalid.", "type": [ "string", "null" ] }, "googleModel": { "description": "Google model.", "enum": [ "chat-bison", "google-gemini-1.5-flash", "google-gemini-1.5-pro" ], "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "googleRegion": { "description": "Google model region.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "googleServiceAccount": { "description": "The ID of the user credential containing a Google service account.", "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "intervention": { "description": "Intervention configuration for the guard.", "properties": { "action": { "description": "The action to take if conditions are met.", "enum": [ "block", "report", "replace" ], "type": "string" }, "allowedActions": { "description": "The actions this guard is allowed to take.", "items": { "enum": [ "block", "report", "replace" ], "type": "string" }, "maxItems": 10, "type": "array" }, "conditionLogic": { "default": "any", "description": "The action to take if conditions are met.", "enum": [ "any" ], "type": "string" }, "conditions": { "description": "The list of conditions to trigger intervention.", "items": { "description": "The condition to trigger intervention.", "properties": { "comparand": { "anyOf": [ { "type": "boolean" }, { "type": "number" }, { "type": "string" }, { "items": { "type": "string" }, "maxItems": 10, "type": "array" } ], "description": "The condition comparand (basis of comparison)." }, "comparator": { "description": "The condition comparator (operator).", "enum": [ "greaterThan", "lessThan", "equals", "notEquals", "is", "isNot", "matches", "doesNotMatch", "contains", "doesNotContain" ], "type": "string" } }, "required": [ "comparand", "comparator" ], "type": "object", "x-versionadded": "v2.35" }, "maxItems": 1, "type": "array" }, "message": { "description": "The message to use if the prompt or response is blocked.", "maxLength": 4096, "type": "string" }, "sendNotification": { "default": false, "description": "The notification event to create if intervention is triggered.", "type": "boolean" } }, "required": [ "action", "conditions", "message" ], "type": "object", "x-versionadded": "v2.35" }, "isAgentic": { "default": false, "description": "True if the guard is suitable for agentic workflows only.", "type": "boolean", "x-versionadded": "v2.37" }, "isValid": { "description": "Whether the guard is valid or not.", "type": "boolean" }, "llmGatewayModelId": { "description": "The LLM Gateway model ID to use as judge.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "llmType": { "description": "The type of LLM used by this guard.", "enum": [ "openAi", "azureOpenAi", "google", "amazon", "datarobot", "nim", "llmGateway" ], "type": [ "string", "null" ] }, "modelInfo": { "description": "The configuration info for guards using deployed models.", "properties": { "classNames": { "description": "The list of class names for multiclass models.", "items": { "description": "The class name.", "maxLength": 128, "type": "string" }, "maxItems": 100, "type": "array" }, "inputColumnName": { "description": "The input column name.", "maxLength": 255, "type": "string" }, "modelId": { "description": "The ID of the registered model for model guards.", "type": [ "string", "null" ] }, "modelName": { "default": "", "description": "The ID of the registered model for .model guards.", "maxLength": 255, "type": "string" }, "outputColumnName": { "description": "The output column name.", "maxLength": 255, "type": "string" }, "replacementTextColumnName": { "default": "", "description": "The name of the output column with replacement text. Required only if intervention.action is `replace`.", "maxLength": 255, "type": "string" }, "targetType": { "description": "The target type.", "enum": [ "Binary", "Regression", "Multiclass", "TextGeneration" ], "type": "string" } }, "required": [ "inputColumnName", "outputColumnName", "targetType" ], "type": "object", "x-versionadded": "v2.35" }, "name": { "description": "Guard configuration name", "maxLength": 255, "type": "string" }, "nemoEvaluatorType": { "description": "Guard configuration \"NeMo Evaluator\" metric type", "enum": [ "llm_judge", "context_relevance", "response_groundedness", "topic_adherence", "agent_goal_accuracy", "response_relevancy", "faithfulness" ], "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "nemoInfo": { "description": "The configuration info for NeMo guards.", "properties": { "actions": { "description": "The NeMo guardrails actions file.", "maxLength": 4096, "type": "string" }, "blockedTerms": { "description": "The NeMo guardrails blocked terms list.", "maxLength": 4096, "type": "string" }, "credentialId": { "description": "The NeMo guardrails credential ID (deprecated; use \"openai_credential\").", "type": [ "string", "null" ] }, "llmPrompts": { "description": "The NeMo guardrails prompts.", "maxLength": 4096, "type": "string" }, "mainConfig": { "description": "The overall NeMo configuration YAML.", "maxLength": 4096, "type": "string" }, "railsConfig": { "description": "The NeMo guardrails configuration Colang.", "maxLength": 4096, "type": "string" } }, "required": [ "blockedTerms", "mainConfig", "railsConfig" ], "type": "object", "x-versionadded": "v2.35" }, "nemoLlmJudgeConfig": { "description": "The configuration for the LLM Judge metric.", "properties": { "customMetricDirectionality": { "description": "The custom metric directionality for the LLM judge.", "enum": [ "higherIsBetter", "lowerIsBetter" ], "type": "string", "x-versionadded": "v2.42" }, "scoreParsingRegex": { "description": "The regex to parse the score from the LLM judge response.", "maxLength": 1024, "type": "string" }, "systemPrompt": { "description": "The system prompt for the LLM judge.", "maxLength": 8192, "type": "string" }, "userPrompt": { "description": "The user prompt template for the LLM judge.", "maxLength": 8192, "type": "string" } }, "required": [ "customMetricDirectionality", "scoreParsingRegex", "systemPrompt", "userPrompt" ], "type": "object", "x-versionadded": "v2.42" }, "nemoResponseRelevancyConfig": { "description": "The configuration for the Response Relevancy metric.", "properties": { "embeddingDeploymentId": { "description": "The ID of the embedding model deployment.", "type": "string" } }, "required": [ "embeddingDeploymentId" ], "type": "object", "x-versionadded": "v2.42" }, "nemoTopicAdherenceConfig": { "description": "The configuration for the Topic Adherence metric.", "properties": { "metricMode": { "default": "f1", "description": "The metric calculation mode.", "enum": [ "f1", "recall", "precision" ], "type": "string" }, "referenceTopics": { "description": "The list of reference topics.", "items": { "description": "The reference topic.", "maxLength": 512, "type": "string" }, "maxItems": 100, "minItems": 1, "type": "array" } }, "required": [ "referenceTopics" ], "type": "object", "x-versionadded": "v2.42" }, "ootbType": { "description": "Guard template \"Out of the Box\" metric type", "enum": [ "token_count", "faithfulness", "rouge_1", "agent_goal_accuracy", "agent_goal_accuracy_with_reference", "cost", "task_adherence", "tool_call_accuracy", "agent_guideline_adherence", "agent_latency", "agent_tokens", "agent_cost", "agentGoalAccuracy", "agentGoalAccuracyWithReference", "taskAdherence", "toolCallAccuracy" ], "type": [ "string", "null" ] }, "openaiApiBase": { "description": "The Azure OpenAI API Base URL.", "maxLength": 255, "type": [ "string", "null" ] }, "openaiApiKey": { "description": "Deprecated; use openai_credential instead.", "maxLength": 255, "type": [ "string", "null" ] }, "openaiCredential": { "description": "The ID of the user credential containing an OpenAI token.", "type": [ "string", "null" ] }, "openaiDeploymentId": { "description": "The OpenAI deployment ID.", "maxLength": 255, "type": [ "string", "null" ] }, "parameters": { "description": "Parameter list, not used, deprecated.", "items": { "maxLength": 1, "type": "string" }, "maxItems": 1, "type": "array" }, "stages": { "description": "The stages where the guard is configured to run.", "items": { "enum": [ "prompt", "response" ], "type": "string" }, "maxItems": 16, "type": "array" }, "type": { "description": "The guard configuration type.", "enum": [ "guardModel", "nemo", "nemoEvaluator", "ootb", "pii", "userModel" ], "type": "string" } }, "required": [ "description", "name", "stages", "type" ], "type": "object", "x-versionadded": "v2.35" } ``` Complete guard configuration to push ### Properties | Name | Type | Required | Restrictions | Description | | --- | --- | --- | --- | --- | | additionalGuardConfig | GuardAdditionalConfig | false | | Additional configuration for the guard. | | allowedActions | [string] | false | maxItems: 10 | The actions this guard is allowed to take. | | awsAccount | string,null | false | | The ID of the user credential containing an AWS account. | | awsModel | string,null | false | | AWS model. | | awsRegion | string,null | false | maxLength: 255 | AWS model region. | | deploymentId | string,null | false | | The ID of the deployed model for model guards. | | description | string | true | maxLength: 4096 | Guard configuration description | | errorMessage | string,null | false | | Error message if the guard configuration is invalid. | | googleModel | string,null | false | | Google model. | | googleRegion | string,null | false | maxLength: 255 | Google model region. | | googleServiceAccount | string,null | false | | The ID of the user credential containing a Google service account. | | intervention | GuardConfigurationInterventionResponse | false | | Intervention configuration for the guard. | | isAgentic | boolean | false | | True if the guard is suitable for agentic workflows only. | | isValid | boolean | false | | Whether the guard is valid or not. | | llmGatewayModelId | string,null | false | maxLength: 255 | The LLM Gateway model ID to use as judge. | | llmType | string,null | false | | The type of LLM used by this guard. | | modelInfo | GuardModelInfoResponse | false | | The configuration info for guards using deployed models. | | name | string | true | maxLength: 255 | Guard configuration name | | nemoEvaluatorType | string,null | false | | Guard configuration "NeMo Evaluator" metric type | | nemoInfo | GuardConfigurationNemoInfoResponse | false | | The configuration info for NeMo guards. | | nemoLlmJudgeConfig | GuardNemoLlmJudgeConfigResponse | false | | The configuration for the LLM Judge metric. | | nemoResponseRelevancyConfig | GuardNemoResponseRelevancyConfigResponse | false | | The configuration for the Response Relevancy metric. | | nemoTopicAdherenceConfig | GuardNemoTopicAdherenceConfigResponse | false | | The configuration for the Topic Adherence metric. | | ootbType | string,null | false | | Guard template "Out of the Box" metric type | | openaiApiBase | string,null | false | maxLength: 255 | The Azure OpenAI API Base URL. | | openaiApiKey | string,null | false | maxLength: 255 | Deprecated; use openai_credential instead. | | openaiCredential | string,null | false | | The ID of the user credential containing an OpenAI token. | | openaiDeploymentId | string,null | false | maxLength: 255 | The OpenAI deployment ID. | | parameters | [string] | false | maxItems: 1 | Parameter list, not used, deprecated. | | stages | [string] | true | maxItems: 16 | The stages where the guard is configured to run. | | type | string | true | | The guard configuration type. | ### Enumerated Values | Property | Value | | --- | --- | | awsModel | [amazon-titan, anthropic-claude-2, anthropic-claude-3-haiku, anthropic-claude-3-sonnet, anthropic-claude-3-opus, anthropic-claude-3.5-sonnet-v1, anthropic-claude-3.5-sonnet-v2, amazon-nova-lite, amazon-nova-micro, amazon-nova-pro] | | googleModel | [chat-bison, google-gemini-1.5-flash, google-gemini-1.5-pro] | | llmType | [openAi, azureOpenAi, google, amazon, datarobot, nim, llmGateway] | | nemoEvaluatorType | [llm_judge, context_relevance, response_groundedness, topic_adherence, agent_goal_accuracy, response_relevancy, faithfulness] | | ootbType | [token_count, faithfulness, rouge_1, agent_goal_accuracy, agent_goal_accuracy_with_reference, cost, task_adherence, tool_call_accuracy, agent_guideline_adherence, agent_latency, agent_tokens, agent_cost, agentGoalAccuracy, agentGoalAccuracyWithReference, taskAdherence, toolCallAccuracy] | | type | [guardModel, nemo, nemoEvaluator, ootb, pii, userModel] | ## GuardConfigurationInterventionResponse ``` { "description": "Intervention configuration for the guard.", "properties": { "action": { "description": "The action to take if conditions are met.", "enum": [ "block", "report", "replace" ], "type": "string" }, "allowedActions": { "description": "The actions this guard is allowed to take.", "items": { "enum": [ "block", "report", "replace" ], "type": "string" }, "maxItems": 10, "type": "array" }, "conditionLogic": { "default": "any", "description": "The action to take if conditions are met.", "enum": [ "any" ], "type": "string" }, "conditions": { "description": "The list of conditions to trigger intervention.", "items": { "description": "The condition to trigger intervention.", "properties": { "comparand": { "anyOf": [ { "type": "boolean" }, { "type": "number" }, { "type": "string" }, { "items": { "type": "string" }, "maxItems": 10, "type": "array" } ], "description": "The condition comparand (basis of comparison)." }, "comparator": { "description": "The condition comparator (operator).", "enum": [ "greaterThan", "lessThan", "equals", "notEquals", "is", "isNot", "matches", "doesNotMatch", "contains", "doesNotContain" ], "type": "string" } }, "required": [ "comparand", "comparator" ], "type": "object", "x-versionadded": "v2.35" }, "maxItems": 1, "type": "array" }, "message": { "description": "The message to use if the prompt or response is blocked.", "maxLength": 4096, "type": "string" }, "sendNotification": { "default": false, "description": "The notification event to create if intervention is triggered.", "type": "boolean" } }, "required": [ "action", "conditions", "message" ], "type": "object", "x-versionadded": "v2.35" } ``` Intervention configuration for the guard. ### Properties | Name | Type | Required | Restrictions | Description | | --- | --- | --- | --- | --- | | action | string | true | | The action to take if conditions are met. | | allowedActions | [string] | false | maxItems: 10 | The actions this guard is allowed to take. | | conditionLogic | string | false | | The action to take if conditions are met. | | conditions | [GuardConfigurationConditionResponse] | true | maxItems: 1 | The list of conditions to trigger intervention. | | message | string | true | maxLength: 4096 | The message to use if the prompt or response is blocked. | | sendNotification | boolean | false | | The notification event to create if intervention is triggered. | ### Enumerated Values | Property | Value | | --- | --- | | action | [block, report, replace] | | conditionLogic | any | ## GuardConfigurationListResponse ``` { "properties": { "count": { "description": "The number of items returned on this page.", "type": "integer" }, "data": { "description": "The list of guard configurations.", "items": { "properties": { "additionalGuardConfig": { "description": "Additional configuration for the guard.", "properties": { "agentGuideline": { "description": "How to calculate agent guideline adherence for this guard.", "maxLength": 4096, "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "cost": { "description": "How to calculate cost information for this guard.", "properties": { "currency": { "description": "ISO 4217 Currency code for display.", "maxLength": 255, "type": "string" }, "inputPrice": { "description": "Cost per unit measure of tokens for input.", "minimum": 0, "type": "number" }, "inputUnit": { "description": "Number of tokens related to input price.", "minimum": 0, "type": "integer" }, "outputPrice": { "description": "Cost per unit measure of tokens for output.", "minimum": 0, "type": "number" }, "outputUnit": { "description": "Number of tokens related to output price.", "minimum": 0, "type": "integer" } }, "required": [ "currency", "inputPrice", "inputUnit", "outputPrice", "outputUnit" ], "type": "object", "x-versionadded": "v2.37" }, "toolCall": { "description": "How to calculate tool call metrics for this guard.", "properties": { "comparisonMetric": { "description": "How tool calls should be compared for metrics calculations.", "enum": [ "exactMatch", "doNotCompare" ], "type": "string" } }, "required": [ "comparisonMetric" ], "type": "object", "x-versionadded": "v2.37" } }, "type": "object", "x-versionadded": "v2.37" }, "allowedActions": { "description": "The actions this guard is allowed to take.", "items": { "enum": [ "block", "report", "replace" ], "type": "string" }, "maxItems": 10, "type": "array", "x-versionadded": "v2.36" }, "awsAccount": { "description": "The ID of the user credential containing an AWS account.", "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "awsModel": { "description": "AWS model.", "enum": [ "amazon-titan", "anthropic-claude-2", "anthropic-claude-3-haiku", "anthropic-claude-3-sonnet", "anthropic-claude-3-opus", "anthropic-claude-3.5-sonnet-v1", "anthropic-claude-3.5-sonnet-v2", "amazon-nova-lite", "amazon-nova-micro", "amazon-nova-pro" ], "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "awsRegion": { "description": "AWS model region.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "createdAt": { "description": "When the configuration was created.", "format": "date-time", "type": "string" }, "creatorId": { "description": "The ID of the user who created the guard configuration.", "type": [ "string", "null" ] }, "creatorName": { "description": "The name of the user who created the guard configuration.", "maxLength": 1000, "type": "string" }, "deploymentId": { "description": "The ID of the deployed model for model guards.", "type": [ "string", "null" ] }, "description": { "description": "The guard configuration description.", "maxLength": 4096, "type": "string" }, "entityId": { "description": "The ID of the custom model or playground for this guard.", "type": [ "string", "null" ] }, "entityType": { "description": "The type of the associated entity.", "enum": [ "customModel", "customModelVersion", "playground" ], "type": "string" }, "errorMessage": { "description": "Error message if the guard configuration is invalid.", "type": [ "string", "null" ] }, "googleModel": { "description": "Google model.", "enum": [ "chat-bison", "google-gemini-1.5-flash", "google-gemini-1.5-pro" ], "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "googleRegion": { "description": "Google model region.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "googleServiceAccount": { "description": "The ID of the user credential containing a Google service account.", "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "id": { "description": "The guard configuration object ID.", "type": "string" }, "intervention": { "description": "Intervention configuration for the guard.", "properties": { "action": { "description": "The action to take if conditions are met.", "enum": [ "block", "report", "replace" ], "type": "string" }, "allowedActions": { "description": "The actions this guard is allowed to take.", "items": { "enum": [ "block", "report", "replace" ], "type": "string" }, "maxItems": 10, "type": "array" }, "conditionLogic": { "default": "any", "description": "The action to take if conditions are met.", "enum": [ "any" ], "type": "string" }, "conditions": { "description": "The list of conditions to trigger intervention.", "items": { "description": "The condition to trigger intervention.", "properties": { "comparand": { "anyOf": [ { "type": "boolean" }, { "type": "number" }, { "type": "string" }, { "items": { "type": "string" }, "maxItems": 10, "type": "array" } ], "description": "The condition comparand (basis of comparison)." }, "comparator": { "description": "The condition comparator (operator).", "enum": [ "greaterThan", "lessThan", "equals", "notEquals", "is", "isNot", "matches", "doesNotMatch", "contains", "doesNotContain" ], "type": "string" } }, "required": [ "comparand", "comparator" ], "type": "object", "x-versionadded": "v2.35" }, "maxItems": 1, "type": "array" }, "message": { "description": "The message to use if the prompt or response is blocked.", "maxLength": 4096, "type": "string" }, "sendNotification": { "default": false, "description": "The notification event to create if intervention is triggered.", "type": "boolean" } }, "required": [ "action", "conditions", "message" ], "type": "object", "x-versionadded": "v2.35" }, "isAgentic": { "default": false, "description": "True if the guard is suitable for agentic workflows only.", "type": "boolean", "x-versionadded": "v2.37" }, "isValid": { "description": "Whether the guard is valid or not.", "type": "boolean" }, "llmGatewayModelId": { "description": "The LLM Gateway model ID to use as judge.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "llmType": { "description": "The type of LLM used by this guard.", "enum": [ "openAi", "azureOpenAi", "google", "amazon", "datarobot", "nim", "llmGateway" ], "type": [ "string", "null" ] }, "modelInfo": { "description": "The configuration info for guards using deployed models.", "properties": { "classNames": { "description": "The list of class names for multiclass models.", "items": { "description": "The class name.", "maxLength": 128, "type": "string" }, "maxItems": 100, "type": "array" }, "inputColumnName": { "description": "The input column name.", "maxLength": 255, "type": "string" }, "modelId": { "description": "The ID of the registered model for model guards.", "type": [ "string", "null" ] }, "modelName": { "default": "", "description": "The ID of the registered model for .model guards.", "maxLength": 255, "type": "string" }, "outputColumnName": { "description": "The output column name.", "maxLength": 255, "type": "string" }, "replacementTextColumnName": { "default": "", "description": "The name of the output column with replacement text. Required only if intervention.action is `replace`.", "maxLength": 255, "type": "string" }, "targetType": { "description": "The target type.", "enum": [ "Binary", "Regression", "Multiclass", "TextGeneration" ], "type": "string" } }, "required": [ "inputColumnName", "outputColumnName", "targetType" ], "type": "object", "x-versionadded": "v2.35" }, "name": { "description": "The guard configuration name.", "maxLength": 255, "type": "string" }, "nemoEvaluatorType": { "description": "Guard configuration \"NeMo Evaluator\" metric type", "enum": [ "llm_judge", "context_relevance", "response_groundedness", "topic_adherence", "agent_goal_accuracy", "response_relevancy", "faithfulness" ], "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "nemoInfo": { "description": "The configuration info for NeMo guards.", "properties": { "actions": { "description": "The NeMo guardrails actions file.", "maxLength": 4096, "type": "string" }, "blockedTerms": { "description": "The NeMo guardrails blocked terms list.", "maxLength": 4096, "type": "string" }, "credentialId": { "description": "The NeMo guardrails credential ID (deprecated; use \"openai_credential\").", "type": [ "string", "null" ] }, "llmPrompts": { "description": "The NeMo guardrails prompts.", "maxLength": 4096, "type": "string" }, "mainConfig": { "description": "The overall NeMo configuration YAML.", "maxLength": 4096, "type": "string" }, "railsConfig": { "description": "The NeMo guardrails configuration Colang.", "maxLength": 4096, "type": "string" } }, "required": [ "blockedTerms", "mainConfig", "railsConfig" ], "type": "object", "x-versionadded": "v2.35" }, "nemoLlmJudgeConfig": { "description": "The configuration for the LLM Judge metric.", "properties": { "customMetricDirectionality": { "description": "The custom metric directionality for the LLM judge.", "enum": [ "higherIsBetter", "lowerIsBetter" ], "type": "string", "x-versionadded": "v2.42" }, "scoreParsingRegex": { "description": "The regex to parse the score from the LLM judge response.", "maxLength": 1024, "type": "string" }, "systemPrompt": { "description": "The system prompt for the LLM judge.", "maxLength": 8192, "type": "string" }, "userPrompt": { "description": "The user prompt template for the LLM judge.", "maxLength": 8192, "type": "string" } }, "required": [ "customMetricDirectionality", "scoreParsingRegex", "systemPrompt", "userPrompt" ], "type": "object", "x-versionadded": "v2.42" }, "nemoResponseRelevancyConfig": { "description": "The configuration for the Response Relevancy metric.", "properties": { "embeddingDeploymentId": { "description": "The ID of the embedding model deployment.", "type": "string" } }, "required": [ "embeddingDeploymentId" ], "type": "object", "x-versionadded": "v2.42" }, "nemoTopicAdherenceConfig": { "description": "The configuration for the Topic Adherence metric.", "properties": { "metricMode": { "default": "f1", "description": "The metric calculation mode.", "enum": [ "f1", "recall", "precision" ], "type": "string" }, "referenceTopics": { "description": "The list of reference topics.", "items": { "description": "The reference topic.", "maxLength": 512, "type": "string" }, "maxItems": 100, "minItems": 1, "type": "array" } }, "required": [ "referenceTopics" ], "type": "object", "x-versionadded": "v2.42" }, "ootbType": { "description": "Guard template \"Out of the Box\" metric type", "enum": [ "token_count", "faithfulness", "rouge_1", "agent_goal_accuracy", "agent_goal_accuracy_with_reference", "cost", "task_adherence", "tool_call_accuracy", "agent_guideline_adherence", "agent_latency", "agent_tokens", "agent_cost", "agentGoalAccuracy", "agentGoalAccuracyWithReference", "taskAdherence", "toolCallAccuracy" ], "type": [ "string", "null" ] }, "openaiApiBase": { "description": "The Azure OpenAI API Base URL.", "maxLength": 255, "type": [ "string", "null" ] }, "openaiApiKey": { "description": "Deprecated; use openai_credential instead.", "maxLength": 255, "type": [ "string", "null" ] }, "openaiCredential": { "description": "The ID of the user credential containing an OpenAI token.", "type": [ "string", "null" ] }, "openaiDeploymentId": { "description": "The OpenAPI deployment ID.", "maxLength": 255, "type": [ "string", "null" ] }, "stages": { "description": "The stages where the guard is configured to run.", "items": { "enum": [ "prompt", "response" ], "type": "string" }, "maxItems": 16, "type": "array" }, "type": { "description": "The guard configuration type.", "enum": [ "guardModel", "nemo", "nemoEvaluator", "ootb", "pii", "userModel" ], "type": "string" } }, "required": [ "createdAt", "description", "entityId", "entityType", "id", "name", "stages", "type" ], "type": "object", "x-versionadded": "v2.35" }, "maxItems": 200, "type": "array" }, "next": { "description": "The URL of the next page (if null, there is no next page).", "format": "uri", "type": [ "string", "null" ] }, "previous": { "description": "The URL of the previous page (if null, there is no previous page).", "format": "uri", "type": [ "string", "null" ] }, "totalCount": { "description": "The total number of items across all pages.", "type": "integer" } }, "required": [ "data", "next", "previous", "totalCount" ], "type": "object", "x-versionadded": "v2.35" } ``` ### Properties | Name | Type | Required | Restrictions | Description | | --- | --- | --- | --- | --- | | count | integer | false | | The number of items returned on this page. | | data | [GuardConfigurationRetrieveResponse] | true | maxItems: 200 | The list of guard configurations. | | next | string,null(uri) | true | | The URL of the next page (if null, there is no next page). | | previous | string,null(uri) | true | | The URL of the previous page (if null, there is no previous page). | | totalCount | integer | true | | The total number of items across all pages. | ## GuardConfigurationNemoInfoResponse ``` { "description": "The configuration info for NeMo guards.", "properties": { "actions": { "description": "The NeMo guardrails actions file.", "maxLength": 4096, "type": "string" }, "blockedTerms": { "description": "The NeMo guardrails blocked terms list.", "maxLength": 4096, "type": "string" }, "credentialId": { "description": "The NeMo guardrails credential ID (deprecated; use \"openai_credential\").", "type": [ "string", "null" ] }, "llmPrompts": { "description": "The NeMo guardrails prompts.", "maxLength": 4096, "type": "string" }, "mainConfig": { "description": "The overall NeMo configuration YAML.", "maxLength": 4096, "type": "string" }, "railsConfig": { "description": "The NeMo guardrails configuration Colang.", "maxLength": 4096, "type": "string" } }, "required": [ "blockedTerms", "mainConfig", "railsConfig" ], "type": "object", "x-versionadded": "v2.35" } ``` The configuration info for NeMo guards. ### Properties | Name | Type | Required | Restrictions | Description | | --- | --- | --- | --- | --- | | actions | string | false | maxLength: 4096 | The NeMo guardrails actions file. | | blockedTerms | string | true | maxLength: 4096 | The NeMo guardrails blocked terms list. | | credentialId | string,null | false | | The NeMo guardrails credential ID (deprecated; use "openai_credential"). | | llmPrompts | string | false | maxLength: 4096 | The NeMo guardrails prompts. | | mainConfig | string | true | maxLength: 4096 | The overall NeMo configuration YAML. | | railsConfig | string | true | maxLength: 4096 | The NeMo guardrails configuration Colang. | ## GuardConfigurationPayloadModelInfo ``` { "description": "The configuration info for guards using deployed models.", "properties": { "classNames": { "description": "The list of class names for multiclass models.", "items": { "description": "The class name.", "maxLength": 128, "type": "string" }, "maxItems": 100, "type": "array" }, "inputColumnName": { "description": "The input column name.", "maxLength": 255, "type": "string" }, "modelId": { "description": "The ID of the registered model for model guards.", "type": [ "string", "null" ] }, "modelName": { "default": "", "description": "The ID of the registered model for .model guards.", "maxLength": 255, "type": "string" }, "outputColumnName": { "description": "The output column name.", "maxLength": 255, "type": "string" }, "replacementTextColumnName": { "default": "", "description": "The name of the output column with replacement text. Required only if intervention.action is `replace`.", "maxLength": 255, "type": "string" }, "targetType": { "description": "The target type.", "enum": [ "Binary", "Regression", "Multiclass", "TextGeneration" ], "type": [ "string", "null" ] } }, "required": [ "inputColumnName", "outputColumnName" ], "type": "object", "x-versionadded": "v2.35" } ``` The configuration info for guards using deployed models. ### Properties | Name | Type | Required | Restrictions | Description | | --- | --- | --- | --- | --- | | classNames | [string] | false | maxItems: 100 | The list of class names for multiclass models. | | inputColumnName | string | true | maxLength: 255 | The input column name. | | modelId | string,null | false | | The ID of the registered model for model guards. | | modelName | string | false | maxLength: 255 | The ID of the registered model for .model guards. | | outputColumnName | string | true | maxLength: 255 | The output column name. | | replacementTextColumnName | string | false | maxLength: 255 | The name of the output column with replacement text. Required only if intervention.action is replace. | | targetType | string,null | false | | The target type. | ### Enumerated Values | Property | Value | | --- | --- | | targetType | [Binary, Regression, Multiclass, TextGeneration] | ## GuardConfigurationPredictionEnvironmentsInUseListResponse ``` { "properties": { "count": { "description": "The number of items returned on this page.", "type": "integer" }, "data": { "description": "list of prediction environments in use for this custom model version.", "items": { "properties": { "id": { "description": "ID of prediction environment.", "type": "string" }, "name": { "description": "Name of prediction environment.", "type": "string" }, "usedBy": { "description": "Guards using this prediction environment.", "items": { "properties": { "configurationId": { "description": "ID of guard configuration.", "type": "string" }, "deploymentId": { "description": "ID of guard model deployment.", "type": "string" }, "name": { "description": "Name of guard configuration.", "type": "string" } }, "required": [ "configurationId", "deploymentId", "name" ], "type": "object", "x-versionadded": "v2.35" }, "maxItems": 32, "type": "array" } }, "required": [ "id", "name", "usedBy" ], "type": "object", "x-versionadded": "v2.35" }, "maxItems": 32, "type": "array" }, "next": { "description": "The URL of the next page (if null, there is no next page).", "format": "uri", "type": [ "string", "null" ] }, "previous": { "description": "The URL of the previous page (if null, there is no previous page).", "format": "uri", "type": [ "string", "null" ] }, "totalCount": { "description": "The total number of items across all pages.", "type": "integer" } }, "required": [ "data", "next", "previous", "totalCount" ], "type": "object", "x-versionadded": "v2.35" } ``` ### Properties | Name | Type | Required | Restrictions | Description | | --- | --- | --- | --- | --- | | count | integer | false | | The number of items returned on this page. | | data | [PredictionEnvironmentInUseResponse] | true | maxItems: 32 | list of prediction environments in use for this custom model version. | | next | string,null(uri) | true | | The URL of the next page (if null, there is no next page). | | previous | string,null(uri) | true | | The URL of the previous page (if null, there is no previous page). | | totalCount | integer | true | | The total number of items across all pages. | ## GuardConfigurationRetrieveResponse ``` { "properties": { "additionalGuardConfig": { "description": "Additional configuration for the guard.", "properties": { "agentGuideline": { "description": "How to calculate agent guideline adherence for this guard.", "maxLength": 4096, "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "cost": { "description": "How to calculate cost information for this guard.", "properties": { "currency": { "description": "ISO 4217 Currency code for display.", "maxLength": 255, "type": "string" }, "inputPrice": { "description": "Cost per unit measure of tokens for input.", "minimum": 0, "type": "number" }, "inputUnit": { "description": "Number of tokens related to input price.", "minimum": 0, "type": "integer" }, "outputPrice": { "description": "Cost per unit measure of tokens for output.", "minimum": 0, "type": "number" }, "outputUnit": { "description": "Number of tokens related to output price.", "minimum": 0, "type": "integer" } }, "required": [ "currency", "inputPrice", "inputUnit", "outputPrice", "outputUnit" ], "type": "object", "x-versionadded": "v2.37" }, "toolCall": { "description": "How to calculate tool call metrics for this guard.", "properties": { "comparisonMetric": { "description": "How tool calls should be compared for metrics calculations.", "enum": [ "exactMatch", "doNotCompare" ], "type": "string" } }, "required": [ "comparisonMetric" ], "type": "object", "x-versionadded": "v2.37" } }, "type": "object", "x-versionadded": "v2.37" }, "allowedActions": { "description": "The actions this guard is allowed to take.", "items": { "enum": [ "block", "report", "replace" ], "type": "string" }, "maxItems": 10, "type": "array", "x-versionadded": "v2.36" }, "awsAccount": { "description": "The ID of the user credential containing an AWS account.", "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "awsModel": { "description": "AWS model.", "enum": [ "amazon-titan", "anthropic-claude-2", "anthropic-claude-3-haiku", "anthropic-claude-3-sonnet", "anthropic-claude-3-opus", "anthropic-claude-3.5-sonnet-v1", "anthropic-claude-3.5-sonnet-v2", "amazon-nova-lite", "amazon-nova-micro", "amazon-nova-pro" ], "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "awsRegion": { "description": "AWS model region.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "createdAt": { "description": "When the configuration was created.", "format": "date-time", "type": "string" }, "creatorId": { "description": "The ID of the user who created the guard configuration.", "type": [ "string", "null" ] }, "creatorName": { "description": "The name of the user who created the guard configuration.", "maxLength": 1000, "type": "string" }, "deploymentId": { "description": "The ID of the deployed model for model guards.", "type": [ "string", "null" ] }, "description": { "description": "The guard configuration description.", "maxLength": 4096, "type": "string" }, "entityId": { "description": "The ID of the custom model or playground for this guard.", "type": [ "string", "null" ] }, "entityType": { "description": "The type of the associated entity.", "enum": [ "customModel", "customModelVersion", "playground" ], "type": "string" }, "errorMessage": { "description": "Error message if the guard configuration is invalid.", "type": [ "string", "null" ] }, "googleModel": { "description": "Google model.", "enum": [ "chat-bison", "google-gemini-1.5-flash", "google-gemini-1.5-pro" ], "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "googleRegion": { "description": "Google model region.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "googleServiceAccount": { "description": "The ID of the user credential containing a Google service account.", "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "id": { "description": "The guard configuration object ID.", "type": "string" }, "intervention": { "description": "Intervention configuration for the guard.", "properties": { "action": { "description": "The action to take if conditions are met.", "enum": [ "block", "report", "replace" ], "type": "string" }, "allowedActions": { "description": "The actions this guard is allowed to take.", "items": { "enum": [ "block", "report", "replace" ], "type": "string" }, "maxItems": 10, "type": "array" }, "conditionLogic": { "default": "any", "description": "The action to take if conditions are met.", "enum": [ "any" ], "type": "string" }, "conditions": { "description": "The list of conditions to trigger intervention.", "items": { "description": "The condition to trigger intervention.", "properties": { "comparand": { "anyOf": [ { "type": "boolean" }, { "type": "number" }, { "type": "string" }, { "items": { "type": "string" }, "maxItems": 10, "type": "array" } ], "description": "The condition comparand (basis of comparison)." }, "comparator": { "description": "The condition comparator (operator).", "enum": [ "greaterThan", "lessThan", "equals", "notEquals", "is", "isNot", "matches", "doesNotMatch", "contains", "doesNotContain" ], "type": "string" } }, "required": [ "comparand", "comparator" ], "type": "object", "x-versionadded": "v2.35" }, "maxItems": 1, "type": "array" }, "message": { "description": "The message to use if the prompt or response is blocked.", "maxLength": 4096, "type": "string" }, "sendNotification": { "default": false, "description": "The notification event to create if intervention is triggered.", "type": "boolean" } }, "required": [ "action", "conditions", "message" ], "type": "object", "x-versionadded": "v2.35" }, "isAgentic": { "default": false, "description": "True if the guard is suitable for agentic workflows only.", "type": "boolean", "x-versionadded": "v2.37" }, "isValid": { "description": "Whether the guard is valid or not.", "type": "boolean" }, "llmGatewayModelId": { "description": "The LLM Gateway model ID to use as judge.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "llmType": { "description": "The type of LLM used by this guard.", "enum": [ "openAi", "azureOpenAi", "google", "amazon", "datarobot", "nim", "llmGateway" ], "type": [ "string", "null" ] }, "modelInfo": { "description": "The configuration info for guards using deployed models.", "properties": { "classNames": { "description": "The list of class names for multiclass models.", "items": { "description": "The class name.", "maxLength": 128, "type": "string" }, "maxItems": 100, "type": "array" }, "inputColumnName": { "description": "The input column name.", "maxLength": 255, "type": "string" }, "modelId": { "description": "The ID of the registered model for model guards.", "type": [ "string", "null" ] }, "modelName": { "default": "", "description": "The ID of the registered model for .model guards.", "maxLength": 255, "type": "string" }, "outputColumnName": { "description": "The output column name.", "maxLength": 255, "type": "string" }, "replacementTextColumnName": { "default": "", "description": "The name of the output column with replacement text. Required only if intervention.action is `replace`.", "maxLength": 255, "type": "string" }, "targetType": { "description": "The target type.", "enum": [ "Binary", "Regression", "Multiclass", "TextGeneration" ], "type": "string" } }, "required": [ "inputColumnName", "outputColumnName", "targetType" ], "type": "object", "x-versionadded": "v2.35" }, "name": { "description": "The guard configuration name.", "maxLength": 255, "type": "string" }, "nemoEvaluatorType": { "description": "Guard configuration \"NeMo Evaluator\" metric type", "enum": [ "llm_judge", "context_relevance", "response_groundedness", "topic_adherence", "agent_goal_accuracy", "response_relevancy", "faithfulness" ], "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "nemoInfo": { "description": "The configuration info for NeMo guards.", "properties": { "actions": { "description": "The NeMo guardrails actions file.", "maxLength": 4096, "type": "string" }, "blockedTerms": { "description": "The NeMo guardrails blocked terms list.", "maxLength": 4096, "type": "string" }, "credentialId": { "description": "The NeMo guardrails credential ID (deprecated; use \"openai_credential\").", "type": [ "string", "null" ] }, "llmPrompts": { "description": "The NeMo guardrails prompts.", "maxLength": 4096, "type": "string" }, "mainConfig": { "description": "The overall NeMo configuration YAML.", "maxLength": 4096, "type": "string" }, "railsConfig": { "description": "The NeMo guardrails configuration Colang.", "maxLength": 4096, "type": "string" } }, "required": [ "blockedTerms", "mainConfig", "railsConfig" ], "type": "object", "x-versionadded": "v2.35" }, "nemoLlmJudgeConfig": { "description": "The configuration for the LLM Judge metric.", "properties": { "customMetricDirectionality": { "description": "The custom metric directionality for the LLM judge.", "enum": [ "higherIsBetter", "lowerIsBetter" ], "type": "string", "x-versionadded": "v2.42" }, "scoreParsingRegex": { "description": "The regex to parse the score from the LLM judge response.", "maxLength": 1024, "type": "string" }, "systemPrompt": { "description": "The system prompt for the LLM judge.", "maxLength": 8192, "type": "string" }, "userPrompt": { "description": "The user prompt template for the LLM judge.", "maxLength": 8192, "type": "string" } }, "required": [ "customMetricDirectionality", "scoreParsingRegex", "systemPrompt", "userPrompt" ], "type": "object", "x-versionadded": "v2.42" }, "nemoResponseRelevancyConfig": { "description": "The configuration for the Response Relevancy metric.", "properties": { "embeddingDeploymentId": { "description": "The ID of the embedding model deployment.", "type": "string" } }, "required": [ "embeddingDeploymentId" ], "type": "object", "x-versionadded": "v2.42" }, "nemoTopicAdherenceConfig": { "description": "The configuration for the Topic Adherence metric.", "properties": { "metricMode": { "default": "f1", "description": "The metric calculation mode.", "enum": [ "f1", "recall", "precision" ], "type": "string" }, "referenceTopics": { "description": "The list of reference topics.", "items": { "description": "The reference topic.", "maxLength": 512, "type": "string" }, "maxItems": 100, "minItems": 1, "type": "array" } }, "required": [ "referenceTopics" ], "type": "object", "x-versionadded": "v2.42" }, "ootbType": { "description": "Guard template \"Out of the Box\" metric type", "enum": [ "token_count", "faithfulness", "rouge_1", "agent_goal_accuracy", "agent_goal_accuracy_with_reference", "cost", "task_adherence", "tool_call_accuracy", "agent_guideline_adherence", "agent_latency", "agent_tokens", "agent_cost", "agentGoalAccuracy", "agentGoalAccuracyWithReference", "taskAdherence", "toolCallAccuracy" ], "type": [ "string", "null" ] }, "openaiApiBase": { "description": "The Azure OpenAI API Base URL.", "maxLength": 255, "type": [ "string", "null" ] }, "openaiApiKey": { "description": "Deprecated; use openai_credential instead.", "maxLength": 255, "type": [ "string", "null" ] }, "openaiCredential": { "description": "The ID of the user credential containing an OpenAI token.", "type": [ "string", "null" ] }, "openaiDeploymentId": { "description": "The OpenAPI deployment ID.", "maxLength": 255, "type": [ "string", "null" ] }, "stages": { "description": "The stages where the guard is configured to run.", "items": { "enum": [ "prompt", "response" ], "type": "string" }, "maxItems": 16, "type": "array" }, "type": { "description": "The guard configuration type.", "enum": [ "guardModel", "nemo", "nemoEvaluator", "ootb", "pii", "userModel" ], "type": "string" } }, "required": [ "createdAt", "description", "entityId", "entityType", "id", "name", "stages", "type" ], "type": "object", "x-versionadded": "v2.35" } ``` ### Properties | Name | Type | Required | Restrictions | Description | | --- | --- | --- | --- | --- | | additionalGuardConfig | GuardAdditionalConfig | false | | Additional configuration for the guard. | | allowedActions | [string] | false | maxItems: 10 | The actions this guard is allowed to take. | | awsAccount | string,null | false | | The ID of the user credential containing an AWS account. | | awsModel | string,null | false | | AWS model. | | awsRegion | string,null | false | maxLength: 255 | AWS model region. | | createdAt | string(date-time) | true | | When the configuration was created. | | creatorId | string,null | false | | The ID of the user who created the guard configuration. | | creatorName | string | false | maxLength: 1000 | The name of the user who created the guard configuration. | | deploymentId | string,null | false | | The ID of the deployed model for model guards. | | description | string | true | maxLength: 4096 | The guard configuration description. | | entityId | string,null | true | | The ID of the custom model or playground for this guard. | | entityType | string | true | | The type of the associated entity. | | errorMessage | string,null | false | | Error message if the guard configuration is invalid. | | googleModel | string,null | false | | Google model. | | googleRegion | string,null | false | maxLength: 255 | Google model region. | | googleServiceAccount | string,null | false | | The ID of the user credential containing a Google service account. | | id | string | true | | The guard configuration object ID. | | intervention | GuardConfigurationInterventionResponse | false | | Intervention configuration for the guard. | | isAgentic | boolean | false | | True if the guard is suitable for agentic workflows only. | | isValid | boolean | false | | Whether the guard is valid or not. | | llmGatewayModelId | string,null | false | maxLength: 255 | The LLM Gateway model ID to use as judge. | | llmType | string,null | false | | The type of LLM used by this guard. | | modelInfo | GuardModelInfoResponse | false | | The configuration info for guards using deployed models. | | name | string | true | maxLength: 255 | The guard configuration name. | | nemoEvaluatorType | string,null | false | | Guard configuration "NeMo Evaluator" metric type | | nemoInfo | GuardConfigurationNemoInfoResponse | false | | The configuration info for NeMo guards. | | nemoLlmJudgeConfig | GuardNemoLlmJudgeConfigResponse | false | | The configuration for the LLM Judge metric. | | nemoResponseRelevancyConfig | GuardNemoResponseRelevancyConfigResponse | false | | The configuration for the Response Relevancy metric. | | nemoTopicAdherenceConfig | GuardNemoTopicAdherenceConfigResponse | false | | The configuration for the Topic Adherence metric. | | ootbType | string,null | false | | Guard template "Out of the Box" metric type | | openaiApiBase | string,null | false | maxLength: 255 | The Azure OpenAI API Base URL. | | openaiApiKey | string,null | false | maxLength: 255 | Deprecated; use openai_credential instead. | | openaiCredential | string,null | false | | The ID of the user credential containing an OpenAI token. | | openaiDeploymentId | string,null | false | maxLength: 255 | The OpenAPI deployment ID. | | stages | [string] | true | maxItems: 16 | The stages where the guard is configured to run. | | type | string | true | | The guard configuration type. | ### Enumerated Values | Property | Value | | --- | --- | | awsModel | [amazon-titan, anthropic-claude-2, anthropic-claude-3-haiku, anthropic-claude-3-sonnet, anthropic-claude-3-opus, anthropic-claude-3.5-sonnet-v1, anthropic-claude-3.5-sonnet-v2, amazon-nova-lite, amazon-nova-micro, amazon-nova-pro] | | entityType | [customModel, customModelVersion, playground] | | googleModel | [chat-bison, google-gemini-1.5-flash, google-gemini-1.5-pro] | | llmType | [openAi, azureOpenAi, google, amazon, datarobot, nim, llmGateway] | | nemoEvaluatorType | [llm_judge, context_relevance, response_groundedness, topic_adherence, agent_goal_accuracy, response_relevancy, faithfulness] | | ootbType | [token_count, faithfulness, rouge_1, agent_goal_accuracy, agent_goal_accuracy_with_reference, cost, task_adherence, tool_call_accuracy, agent_guideline_adherence, agent_latency, agent_tokens, agent_cost, agentGoalAccuracy, agentGoalAccuracyWithReference, taskAdherence, toolCallAccuracy] | | type | [guardModel, nemo, nemoEvaluator, ootb, pii, userModel] | ## GuardConfigurationToCustomModelResponse ``` { "properties": { "customModelVersionId": { "description": "The ID of the new custom model version created.", "type": "string" } }, "required": [ "customModelVersionId" ], "type": "object", "x-versionadded": "v2.35" } ``` ### Properties | Name | Type | Required | Restrictions | Description | | --- | --- | --- | --- | --- | | customModelVersionId | string | true | | The ID of the new custom model version created. | ## GuardConfigurationToCustomModelVersion ``` { "properties": { "customModelId": { "description": "The ID of the custom model the user is working with.", "type": "string" }, "data": { "description": "The list of complete guard configurations to push.", "items": { "description": "Complete guard configuration to push", "properties": { "additionalGuardConfig": { "description": "Additional configuration for the guard.", "properties": { "agentGuideline": { "description": "How to calculate agent guideline adherence for this guard.", "maxLength": 4096, "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "cost": { "description": "How to calculate cost information for this guard.", "properties": { "currency": { "description": "ISO 4217 Currency code for display.", "maxLength": 255, "type": "string" }, "inputPrice": { "description": "Cost per unit measure of tokens for input.", "minimum": 0, "type": "number" }, "inputUnit": { "description": "Number of tokens related to input price.", "minimum": 0, "type": "integer" }, "outputPrice": { "description": "Cost per unit measure of tokens for output.", "minimum": 0, "type": "number" }, "outputUnit": { "description": "Number of tokens related to output price.", "minimum": 0, "type": "integer" } }, "required": [ "currency", "inputPrice", "inputUnit", "outputPrice", "outputUnit" ], "type": "object", "x-versionadded": "v2.37" }, "toolCall": { "description": "How to calculate tool call metrics for this guard.", "properties": { "comparisonMetric": { "description": "How tool calls should be compared for metrics calculations.", "enum": [ "exactMatch", "doNotCompare" ], "type": "string" } }, "required": [ "comparisonMetric" ], "type": "object", "x-versionadded": "v2.37" } }, "type": "object", "x-versionadded": "v2.37" }, "allowedActions": { "description": "The actions this guard is allowed to take.", "items": { "enum": [ "block", "report", "replace" ], "type": "string" }, "maxItems": 10, "type": "array", "x-versionadded": "v2.36" }, "awsAccount": { "description": "The ID of the user credential containing an AWS account.", "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "awsModel": { "description": "AWS model.", "enum": [ "amazon-titan", "anthropic-claude-2", "anthropic-claude-3-haiku", "anthropic-claude-3-sonnet", "anthropic-claude-3-opus", "anthropic-claude-3.5-sonnet-v1", "anthropic-claude-3.5-sonnet-v2", "amazon-nova-lite", "amazon-nova-micro", "amazon-nova-pro" ], "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "awsRegion": { "description": "AWS model region.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "deploymentId": { "description": "The ID of the deployed model for model guards.", "type": [ "string", "null" ] }, "description": { "description": "Guard configuration description", "maxLength": 4096, "type": "string" }, "errorMessage": { "description": "Error message if the guard configuration is invalid.", "type": [ "string", "null" ] }, "googleModel": { "description": "Google model.", "enum": [ "chat-bison", "google-gemini-1.5-flash", "google-gemini-1.5-pro" ], "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "googleRegion": { "description": "Google model region.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "googleServiceAccount": { "description": "The ID of the user credential containing a Google service account.", "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "intervention": { "description": "Intervention configuration for the guard.", "properties": { "action": { "description": "The action to take if conditions are met.", "enum": [ "block", "report", "replace" ], "type": "string" }, "allowedActions": { "description": "The actions this guard is allowed to take.", "items": { "enum": [ "block", "report", "replace" ], "type": "string" }, "maxItems": 10, "type": "array" }, "conditionLogic": { "default": "any", "description": "The action to take if conditions are met.", "enum": [ "any" ], "type": "string" }, "conditions": { "description": "The list of conditions to trigger intervention.", "items": { "description": "The condition to trigger intervention.", "properties": { "comparand": { "anyOf": [ { "type": "boolean" }, { "type": "number" }, { "type": "string" }, { "items": { "type": "string" }, "maxItems": 10, "type": "array" } ], "description": "The condition comparand (basis of comparison)." }, "comparator": { "description": "The condition comparator (operator).", "enum": [ "greaterThan", "lessThan", "equals", "notEquals", "is", "isNot", "matches", "doesNotMatch", "contains", "doesNotContain" ], "type": "string" } }, "required": [ "comparand", "comparator" ], "type": "object", "x-versionadded": "v2.35" }, "maxItems": 1, "type": "array" }, "message": { "description": "The message to use if the prompt or response is blocked.", "maxLength": 4096, "type": "string" }, "sendNotification": { "default": false, "description": "The notification event to create if intervention is triggered.", "type": "boolean" } }, "required": [ "action", "conditions", "message" ], "type": "object", "x-versionadded": "v2.35" }, "isAgentic": { "default": false, "description": "True if the guard is suitable for agentic workflows only.", "type": "boolean", "x-versionadded": "v2.37" }, "isValid": { "description": "Whether the guard is valid or not.", "type": "boolean" }, "llmGatewayModelId": { "description": "The LLM Gateway model ID to use as judge.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "llmType": { "description": "The type of LLM used by this guard.", "enum": [ "openAi", "azureOpenAi", "google", "amazon", "datarobot", "nim", "llmGateway" ], "type": [ "string", "null" ] }, "modelInfo": { "description": "The configuration info for guards using deployed models.", "properties": { "classNames": { "description": "The list of class names for multiclass models.", "items": { "description": "The class name.", "maxLength": 128, "type": "string" }, "maxItems": 100, "type": "array" }, "inputColumnName": { "description": "The input column name.", "maxLength": 255, "type": "string" }, "modelId": { "description": "The ID of the registered model for model guards.", "type": [ "string", "null" ] }, "modelName": { "default": "", "description": "The ID of the registered model for .model guards.", "maxLength": 255, "type": "string" }, "outputColumnName": { "description": "The output column name.", "maxLength": 255, "type": "string" }, "replacementTextColumnName": { "default": "", "description": "The name of the output column with replacement text. Required only if intervention.action is `replace`.", "maxLength": 255, "type": "string" }, "targetType": { "description": "The target type.", "enum": [ "Binary", "Regression", "Multiclass", "TextGeneration" ], "type": "string" } }, "required": [ "inputColumnName", "outputColumnName", "targetType" ], "type": "object", "x-versionadded": "v2.35" }, "name": { "description": "Guard configuration name", "maxLength": 255, "type": "string" }, "nemoEvaluatorType": { "description": "Guard configuration \"NeMo Evaluator\" metric type", "enum": [ "llm_judge", "context_relevance", "response_groundedness", "topic_adherence", "agent_goal_accuracy", "response_relevancy", "faithfulness" ], "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "nemoInfo": { "description": "The configuration info for NeMo guards.", "properties": { "actions": { "description": "The NeMo guardrails actions file.", "maxLength": 4096, "type": "string" }, "blockedTerms": { "description": "The NeMo guardrails blocked terms list.", "maxLength": 4096, "type": "string" }, "credentialId": { "description": "The NeMo guardrails credential ID (deprecated; use \"openai_credential\").", "type": [ "string", "null" ] }, "llmPrompts": { "description": "The NeMo guardrails prompts.", "maxLength": 4096, "type": "string" }, "mainConfig": { "description": "The overall NeMo configuration YAML.", "maxLength": 4096, "type": "string" }, "railsConfig": { "description": "The NeMo guardrails configuration Colang.", "maxLength": 4096, "type": "string" } }, "required": [ "blockedTerms", "mainConfig", "railsConfig" ], "type": "object", "x-versionadded": "v2.35" }, "nemoLlmJudgeConfig": { "description": "The configuration for the LLM Judge metric.", "properties": { "customMetricDirectionality": { "description": "The custom metric directionality for the LLM judge.", "enum": [ "higherIsBetter", "lowerIsBetter" ], "type": "string", "x-versionadded": "v2.42" }, "scoreParsingRegex": { "description": "The regex to parse the score from the LLM judge response.", "maxLength": 1024, "type": "string" }, "systemPrompt": { "description": "The system prompt for the LLM judge.", "maxLength": 8192, "type": "string" }, "userPrompt": { "description": "The user prompt template for the LLM judge.", "maxLength": 8192, "type": "string" } }, "required": [ "customMetricDirectionality", "scoreParsingRegex", "systemPrompt", "userPrompt" ], "type": "object", "x-versionadded": "v2.42" }, "nemoResponseRelevancyConfig": { "description": "The configuration for the Response Relevancy metric.", "properties": { "embeddingDeploymentId": { "description": "The ID of the embedding model deployment.", "type": "string" } }, "required": [ "embeddingDeploymentId" ], "type": "object", "x-versionadded": "v2.42" }, "nemoTopicAdherenceConfig": { "description": "The configuration for the Topic Adherence metric.", "properties": { "metricMode": { "default": "f1", "description": "The metric calculation mode.", "enum": [ "f1", "recall", "precision" ], "type": "string" }, "referenceTopics": { "description": "The list of reference topics.", "items": { "description": "The reference topic.", "maxLength": 512, "type": "string" }, "maxItems": 100, "minItems": 1, "type": "array" } }, "required": [ "referenceTopics" ], "type": "object", "x-versionadded": "v2.42" }, "ootbType": { "description": "Guard template \"Out of the Box\" metric type", "enum": [ "token_count", "faithfulness", "rouge_1", "agent_goal_accuracy", "agent_goal_accuracy_with_reference", "cost", "task_adherence", "tool_call_accuracy", "agent_guideline_adherence", "agent_latency", "agent_tokens", "agent_cost", "agentGoalAccuracy", "agentGoalAccuracyWithReference", "taskAdherence", "toolCallAccuracy" ], "type": [ "string", "null" ] }, "openaiApiBase": { "description": "The Azure OpenAI API Base URL.", "maxLength": 255, "type": [ "string", "null" ] }, "openaiApiKey": { "description": "Deprecated; use openai_credential instead.", "maxLength": 255, "type": [ "string", "null" ] }, "openaiCredential": { "description": "The ID of the user credential containing an OpenAI token.", "type": [ "string", "null" ] }, "openaiDeploymentId": { "description": "The OpenAI deployment ID.", "maxLength": 255, "type": [ "string", "null" ] }, "parameters": { "description": "Parameter list, not used, deprecated.", "items": { "maxLength": 1, "type": "string" }, "maxItems": 1, "type": "array" }, "stages": { "description": "The stages where the guard is configured to run.", "items": { "enum": [ "prompt", "response" ], "type": "string" }, "maxItems": 16, "type": "array" }, "type": { "description": "The guard configuration type.", "enum": [ "guardModel", "nemo", "nemoEvaluator", "ootb", "pii", "userModel" ], "type": "string" } }, "required": [ "description", "name", "stages", "type" ], "type": "object", "x-versionadded": "v2.35" }, "maxItems": 200, "type": "array" }, "overallConfig": { "description": "Overall moderation configuration to push (not specific to one guard)", "properties": { "nemoEvaluatorDeploymentId": { "description": "ID of NeMo Evaluator deployment to use for all NeMo Evaluator guards.", "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "timeoutAction": { "description": "Action to take if timeout occurs", "enum": [ "block", "score" ], "type": "string" }, "timeoutSec": { "description": "Timeout value in seconds for any guard", "minimum": 2, "type": "integer" } }, "required": [ "timeoutAction", "timeoutSec" ], "type": "object", "x-versionadded": "v2.35" } }, "required": [ "customModelId", "data" ], "type": "object", "x-versionadded": "v2.35" } ``` ### Properties | Name | Type | Required | Restrictions | Description | | --- | --- | --- | --- | --- | | customModelId | string | true | | The ID of the custom model the user is working with. | | data | [GuardConfigurationFullPost] | true | maxItems: 200 | The list of complete guard configurations to push. | | overallConfig | OverallConfigUpdate | false | | Overall moderation configuration to push (not specific to one guard) | ## GuardConfigurationUpdate ``` { "properties": { "additionalGuardConfig": { "description": "Additional configuration for the guard.", "properties": { "agentGuideline": { "description": "How to calculate agent guideline adherence for this guard.", "maxLength": 4096, "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "cost": { "description": "How to calculate cost information for this guard.", "properties": { "currency": { "description": "ISO 4217 Currency code for display.", "maxLength": 255, "type": "string" }, "inputPrice": { "description": "Cost per unit measure of tokens for input.", "minimum": 0, "type": "number" }, "inputUnit": { "description": "Number of tokens related to input price.", "minimum": 0, "type": "integer" }, "outputPrice": { "description": "Cost per unit measure of tokens for output.", "minimum": 0, "type": "number" }, "outputUnit": { "description": "Number of tokens related to output price.", "minimum": 0, "type": "integer" } }, "required": [ "currency", "inputPrice", "inputUnit", "outputPrice", "outputUnit" ], "type": "object", "x-versionadded": "v2.37" }, "toolCall": { "description": "How to calculate tool call metrics for this guard.", "properties": { "comparisonMetric": { "description": "How tool calls should be compared for metrics calculations.", "enum": [ "exactMatch", "doNotCompare" ], "type": "string" } }, "required": [ "comparisonMetric" ], "type": "object", "x-versionadded": "v2.37" } }, "type": "object", "x-versionadded": "v2.37" }, "allowedActions": { "description": "The actions this guard is allowed to take.", "items": { "enum": [ "block", "report", "replace" ], "type": "string" }, "maxItems": 10, "type": "array", "x-versionadded": "v2.36" }, "awsAccount": { "description": "The ID of the user credential containing an AWS account.", "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "awsModel": { "description": "AWS model.", "enum": [ "amazon-titan", "anthropic-claude-2", "anthropic-claude-3-haiku", "anthropic-claude-3-sonnet", "anthropic-claude-3-opus", "anthropic-claude-3.5-sonnet-v1", "anthropic-claude-3.5-sonnet-v2", "amazon-nova-lite", "amazon-nova-micro", "amazon-nova-pro" ], "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "awsRegion": { "description": "AWS model region.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "deploymentId": { "description": "The ID of the deployed model for model guards.", "type": [ "string", "null" ] }, "description": { "description": "Guard configuration description", "maxLength": 4096, "type": "string" }, "googleModel": { "description": "Google model.", "enum": [ "chat-bison", "google-gemini-1.5-flash", "google-gemini-1.5-pro" ], "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "googleRegion": { "description": "Google model region.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "googleServiceAccount": { "description": "The ID of the user credential containing a Google service account.", "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "intervention": { "description": "Intervention configuration for the guard.", "properties": { "action": { "description": "The action to take if conditions are met.", "enum": [ "block", "report", "replace" ], "type": "string" }, "allowedActions": { "description": "The actions this guard is allowed to take.", "items": { "enum": [ "block", "report", "replace" ], "type": "string" }, "maxItems": 10, "type": "array" }, "conditionLogic": { "default": "any", "description": "The action to take if conditions are met.", "enum": [ "any" ], "type": "string" }, "conditions": { "description": "The list of conditions to trigger intervention.", "items": { "description": "The condition to trigger intervention.", "properties": { "comparand": { "anyOf": [ { "type": "boolean" }, { "type": "number" }, { "type": "string" }, { "items": { "type": "string" }, "maxItems": 10, "type": "array" } ], "description": "The condition comparand (basis of comparison)." }, "comparator": { "description": "The condition comparator (operator).", "enum": [ "greaterThan", "lessThan", "equals", "notEquals", "is", "isNot", "matches", "doesNotMatch", "contains", "doesNotContain" ], "type": "string" } }, "required": [ "comparand", "comparator" ], "type": "object", "x-versionadded": "v2.35" }, "maxItems": 1, "type": "array" }, "message": { "description": "The message to use if the prompt or response is blocked.", "maxLength": 4096, "type": "string" }, "sendNotification": { "default": false, "description": "The notification event to create if intervention is triggered.", "type": "boolean" } }, "required": [ "action", "conditions", "message" ], "type": "object", "x-versionadded": "v2.35" }, "llmGatewayModelId": { "description": "The LLM Gateway model ID to use as judge.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "llmType": { "description": "The type of LLM used by this guard.", "enum": [ "openAi", "azureOpenAi", "google", "amazon", "datarobot", "nim", "llmGateway" ], "type": [ "string", "null" ] }, "modelInfo": { "description": "The configuration info for guards using deployed models.", "properties": { "classNames": { "description": "The list of class names for multiclass models.", "items": { "description": "The class name.", "maxLength": 128, "type": "string" }, "maxItems": 100, "type": "array" }, "inputColumnName": { "description": "The input column name.", "maxLength": 255, "type": "string" }, "modelId": { "description": "The ID of the registered model for model guards.", "type": [ "string", "null" ] }, "modelName": { "default": "", "description": "The ID of the registered model for .model guards.", "maxLength": 255, "type": "string" }, "outputColumnName": { "description": "The output column name.", "maxLength": 255, "type": "string" }, "replacementTextColumnName": { "default": "", "description": "The name of the output column with replacement text. Required only if intervention.action is `replace`.", "maxLength": 255, "type": "string" }, "targetType": { "description": "The target type.", "enum": [ "Binary", "Regression", "Multiclass", "TextGeneration" ], "type": [ "string", "null" ] } }, "required": [ "inputColumnName", "outputColumnName" ], "type": "object", "x-versionadded": "v2.35" }, "name": { "description": "Guard configuration name", "maxLength": 255, "type": "string" }, "nemoInfo": { "description": "The configuration info for NeMo guards.", "properties": { "actions": { "description": "The NeMo guardrails actions file.", "maxLength": 4096, "type": "string" }, "blockedTerms": { "description": "The NeMo guardrails blocked terms list.", "maxLength": 4096, "type": "string" }, "credentialId": { "description": "The NeMo guardrails credential ID (deprecated; use \"openai_credential\").", "type": [ "string", "null" ] }, "llmPrompts": { "description": "The NeMo guardrails prompts.", "maxLength": 4096, "type": "string" }, "mainConfig": { "description": "The overall NeMo configuration YAML.", "maxLength": 4096, "type": "string" }, "railsConfig": { "description": "The NeMo guardrails configuration Colang.", "maxLength": 4096, "type": "string" } }, "required": [ "blockedTerms", "mainConfig", "railsConfig" ], "type": "object", "x-versionadded": "v2.35" }, "nemoLlmJudgeConfig": { "description": "The configuration for the LLM Judge metric.", "properties": { "customMetricDirectionality": { "description": "The custom metric directionality for the LLM judge.", "enum": [ "higherIsBetter", "lowerIsBetter" ], "type": "string", "x-versionadded": "v2.42" }, "scoreParsingRegex": { "description": "The regex to parse the score from the LLM judge response.", "maxLength": 1024, "type": "string" }, "systemPrompt": { "description": "The system prompt for the LLM judge.", "maxLength": 8192, "type": "string" }, "userPrompt": { "description": "The user prompt template for the LLM judge.", "maxLength": 8192, "type": "string" } }, "required": [ "customMetricDirectionality", "scoreParsingRegex", "systemPrompt", "userPrompt" ], "type": "object", "x-versionadded": "v2.42" }, "nemoResponseRelevancyConfig": { "description": "The configuration for the Response Relevancy metric.", "properties": { "embeddingDeploymentId": { "description": "The ID of the embedding model deployment.", "type": "string" } }, "required": [ "embeddingDeploymentId" ], "type": "object", "x-versionadded": "v2.42" }, "nemoTopicAdherenceConfig": { "description": "The configuration for the Topic Adherence metric.", "properties": { "metricMode": { "default": "f1", "description": "The metric calculation mode.", "enum": [ "f1", "recall", "precision" ], "type": "string" }, "referenceTopics": { "description": "The list of reference topics.", "items": { "description": "The reference topic.", "maxLength": 512, "type": "string" }, "maxItems": 100, "minItems": 1, "type": "array" } }, "required": [ "referenceTopics" ], "type": "object", "x-versionadded": "v2.42" }, "openaiApiBase": { "description": "The Azure OpenAI API Base URL.", "maxLength": 255, "type": [ "string", "null" ] }, "openaiApiKey": { "description": "Deprecated; use openai_credential instead.", "maxLength": 255, "type": [ "string", "null" ] }, "openaiCredential": { "description": "The ID of the user credential containing an OpenAI token.", "type": [ "string", "null" ] }, "openaiDeploymentId": { "description": "The OpenAPI deployment ID.", "maxLength": 255, "type": [ "string", "null" ] } }, "type": "object", "x-versionadded": "v2.35" } ``` ### Properties | Name | Type | Required | Restrictions | Description | | --- | --- | --- | --- | --- | | additionalGuardConfig | GuardAdditionalConfig | false | | Additional configuration for the guard. | | allowedActions | [string] | false | maxItems: 10 | The actions this guard is allowed to take. | | awsAccount | string,null | false | | The ID of the user credential containing an AWS account. | | awsModel | string,null | false | | AWS model. | | awsRegion | string,null | false | maxLength: 255 | AWS model region. | | deploymentId | string,null | false | | The ID of the deployed model for model guards. | | description | string | false | maxLength: 4096 | Guard configuration description | | googleModel | string,null | false | | Google model. | | googleRegion | string,null | false | maxLength: 255 | Google model region. | | googleServiceAccount | string,null | false | | The ID of the user credential containing a Google service account. | | intervention | GuardConfigurationInterventionResponse | false | | Intervention configuration for the guard. | | llmGatewayModelId | string,null | false | maxLength: 255 | The LLM Gateway model ID to use as judge. | | llmType | string,null | false | | The type of LLM used by this guard. | | modelInfo | GuardConfigurationPayloadModelInfo | false | | The configuration info for guards using deployed models. | | name | string | false | maxLength: 255 | Guard configuration name | | nemoInfo | GuardConfigurationNemoInfoResponse | false | | The configuration info for NeMo guards. | | nemoLlmJudgeConfig | GuardNemoLlmJudgeConfigResponse | false | | The configuration for the LLM Judge metric. | | nemoResponseRelevancyConfig | GuardNemoResponseRelevancyConfigResponse | false | | The configuration for the Response Relevancy metric. | | nemoTopicAdherenceConfig | GuardNemoTopicAdherenceConfigResponse | false | | The configuration for the Topic Adherence metric. | | openaiApiBase | string,null | false | maxLength: 255 | The Azure OpenAI API Base URL. | | openaiApiKey | string,null | false | maxLength: 255 | Deprecated; use openai_credential instead. | | openaiCredential | string,null | false | | The ID of the user credential containing an OpenAI token. | | openaiDeploymentId | string,null | false | maxLength: 255 | The OpenAPI deployment ID. | ### Enumerated Values | Property | Value | | --- | --- | | awsModel | [amazon-titan, anthropic-claude-2, anthropic-claude-3-haiku, anthropic-claude-3-sonnet, anthropic-claude-3-opus, anthropic-claude-3.5-sonnet-v1, anthropic-claude-3.5-sonnet-v2, amazon-nova-lite, amazon-nova-micro, amazon-nova-pro] | | googleModel | [chat-bison, google-gemini-1.5-flash, google-gemini-1.5-pro] | | llmType | [openAi, azureOpenAi, google, amazon, datarobot, nim, llmGateway] | ## GuardInterventionResponse ``` { "description": "The intervention configuration for the guard.", "properties": { "action": { "description": "The action to take if conditions are met.", "enum": [ "block", "report", "replace" ], "type": "string" }, "allowedActions": { "description": "The actions this guard is allowed to take.", "items": { "enum": [ "block", "report", "replace" ], "type": "string" }, "maxItems": 10, "type": "array" }, "conditionLogic": { "default": "any", "description": "The action to take if conditions are met.", "enum": [ "any" ], "type": "string" }, "conditions": { "description": "The list of conditions to trigger intervention.", "items": { "description": "The condition to trigger intervention.", "properties": { "comparand": { "anyOf": [ { "type": "boolean" }, { "type": "number" }, { "type": "string" }, { "items": { "description": "The class name to match.", "maxLength": 128, "type": "string" }, "maxItems": 10, "type": "array" } ], "description": "The condition comparand (basis of comparison)." }, "comparator": { "description": "The condition comparator (operator).", "enum": [ "greaterThan", "lessThan", "equals", "notEquals", "is", "isNot", "matches", "doesNotMatch", "contains", "doesNotContain" ], "type": "string" } }, "required": [ "comparand", "comparator" ], "type": "object", "x-versionadded": "v2.35" }, "maxItems": 1, "type": "array" }, "modifyMessage": { "description": "The message to use if the prompt or response is blocked.", "maxLength": 4096, "type": "string" }, "sendNotification": { "description": "The notification event to create if intervention is triggered.", "type": "boolean" } }, "required": [ "action", "conditions", "modifyMessage" ], "type": "object", "x-versionadded": "v2.35" } ``` The intervention configuration for the guard. ### Properties | Name | Type | Required | Restrictions | Description | | --- | --- | --- | --- | --- | | action | string | true | | The action to take if conditions are met. | | allowedActions | [string] | false | maxItems: 10 | The actions this guard is allowed to take. | | conditionLogic | string | false | | The action to take if conditions are met. | | conditions | [GuardConditionResponse] | true | maxItems: 1 | The list of conditions to trigger intervention. | | modifyMessage | string | true | maxLength: 4096 | The message to use if the prompt or response is blocked. | | sendNotification | boolean | false | | The notification event to create if intervention is triggered. | ### Enumerated Values | Property | Value | | --- | --- | | action | [block, report, replace] | | conditionLogic | any | ## GuardModelInfoResponse ``` { "description": "The configuration info for guards using deployed models.", "properties": { "classNames": { "description": "The list of class names for multiclass models.", "items": { "description": "The class name.", "maxLength": 128, "type": "string" }, "maxItems": 100, "type": "array" }, "inputColumnName": { "description": "The input column name.", "maxLength": 255, "type": "string" }, "modelId": { "description": "The ID of the registered model for model guards.", "type": [ "string", "null" ] }, "modelName": { "default": "", "description": "The ID of the registered model for .model guards.", "maxLength": 255, "type": "string" }, "outputColumnName": { "description": "The output column name.", "maxLength": 255, "type": "string" }, "replacementTextColumnName": { "default": "", "description": "The name of the output column with replacement text. Required only if intervention.action is `replace`.", "maxLength": 255, "type": "string" }, "targetType": { "description": "The target type.", "enum": [ "Binary", "Regression", "Multiclass", "TextGeneration" ], "type": "string" } }, "required": [ "inputColumnName", "outputColumnName", "targetType" ], "type": "object", "x-versionadded": "v2.35" } ``` The configuration info for guards using deployed models. ### Properties | Name | Type | Required | Restrictions | Description | | --- | --- | --- | --- | --- | | classNames | [string] | false | maxItems: 100 | The list of class names for multiclass models. | | inputColumnName | string | true | maxLength: 255 | The input column name. | | modelId | string,null | false | | The ID of the registered model for model guards. | | modelName | string | false | maxLength: 255 | The ID of the registered model for .model guards. | | outputColumnName | string | true | maxLength: 255 | The output column name. | | replacementTextColumnName | string | false | maxLength: 255 | The name of the output column with replacement text. Required only if intervention.action is replace. | | targetType | string | true | | The target type. | ### Enumerated Values | Property | Value | | --- | --- | | targetType | [Binary, Regression, Multiclass, TextGeneration] | ## GuardNemoInfoResponse ``` { "description": "The configuration info for NeMo guards.", "properties": { "actions": { "default": "", "description": "The NeMo guardrails actions.", "maxLength": 4096, "type": "string" }, "blockedTerms": { "description": "The NeMo guardrails blocked terms list.", "maxLength": 4096, "type": "string" }, "credentialId": { "description": "The NeMo guardrails credential ID.", "type": [ "string", "null" ] }, "llmPrompts": { "default": "", "description": "The NeMo guardrails prompts.", "maxLength": 4096, "type": "string" }, "mainConfig": { "description": "The overall NeMo configuration YAML.", "maxLength": 4096, "type": "string" }, "railsConfig": { "description": "The NeMo guardrails configuration Colang.", "maxLength": 4096, "type": "string" } }, "required": [ "blockedTerms", "mainConfig", "railsConfig" ], "type": "object", "x-versionadded": "v2.35" } ``` The configuration info for NeMo guards. ### Properties | Name | Type | Required | Restrictions | Description | | --- | --- | --- | --- | --- | | actions | string | false | maxLength: 4096 | The NeMo guardrails actions. | | blockedTerms | string | true | maxLength: 4096 | The NeMo guardrails blocked terms list. | | credentialId | string,null | false | | The NeMo guardrails credential ID. | | llmPrompts | string | false | maxLength: 4096 | The NeMo guardrails prompts. | | mainConfig | string | true | maxLength: 4096 | The overall NeMo configuration YAML. | | railsConfig | string | true | maxLength: 4096 | The NeMo guardrails configuration Colang. | ## GuardNemoLlmJudgeConfigResponse ``` { "description": "The configuration for the LLM Judge metric.", "properties": { "customMetricDirectionality": { "description": "The custom metric directionality for the LLM judge.", "enum": [ "higherIsBetter", "lowerIsBetter" ], "type": "string", "x-versionadded": "v2.42" }, "scoreParsingRegex": { "description": "The regex to parse the score from the LLM judge response.", "maxLength": 1024, "type": "string" }, "systemPrompt": { "description": "The system prompt for the LLM judge.", "maxLength": 8192, "type": "string" }, "userPrompt": { "description": "The user prompt template for the LLM judge.", "maxLength": 8192, "type": "string" } }, "required": [ "customMetricDirectionality", "scoreParsingRegex", "systemPrompt", "userPrompt" ], "type": "object", "x-versionadded": "v2.42" } ``` The configuration for the LLM Judge metric. ### Properties | Name | Type | Required | Restrictions | Description | | --- | --- | --- | --- | --- | | customMetricDirectionality | string | true | | The custom metric directionality for the LLM judge. | | scoreParsingRegex | string | true | maxLength: 1024 | The regex to parse the score from the LLM judge response. | | systemPrompt | string | true | maxLength: 8192 | The system prompt for the LLM judge. | | userPrompt | string | true | maxLength: 8192 | The user prompt template for the LLM judge. | ### Enumerated Values | Property | Value | | --- | --- | | customMetricDirectionality | [higherIsBetter, lowerIsBetter] | ## GuardNemoResponseRelevancyConfigResponse ``` { "description": "The configuration for the Response Relevancy metric.", "properties": { "embeddingDeploymentId": { "description": "The ID of the embedding model deployment.", "type": "string" } }, "required": [ "embeddingDeploymentId" ], "type": "object", "x-versionadded": "v2.42" } ``` The configuration for the Response Relevancy metric. ### Properties | Name | Type | Required | Restrictions | Description | | --- | --- | --- | --- | --- | | embeddingDeploymentId | string | true | | The ID of the embedding model deployment. | ## GuardNemoTopicAdherenceConfigResponse ``` { "description": "The configuration for the Topic Adherence metric.", "properties": { "metricMode": { "default": "f1", "description": "The metric calculation mode.", "enum": [ "f1", "recall", "precision" ], "type": "string" }, "referenceTopics": { "description": "The list of reference topics.", "items": { "description": "The reference topic.", "maxLength": 512, "type": "string" }, "maxItems": 100, "minItems": 1, "type": "array" } }, "required": [ "referenceTopics" ], "type": "object", "x-versionadded": "v2.42" } ``` The configuration for the Topic Adherence metric. ### Properties | Name | Type | Required | Restrictions | Description | | --- | --- | --- | --- | --- | | metricMode | string | false | | The metric calculation mode. | | referenceTopics | [string] | true | maxItems: 100minItems: 1 | The list of reference topics. | ### Enumerated Values | Property | Value | | --- | --- | | metricMode | [f1, recall, precision] | ## GuardTemplateListResponse ``` { "properties": { "count": { "description": "The number of items returned on this page.", "type": "integer" }, "data": { "description": "The list of guard templates.", "items": { "properties": { "allowedActions": { "description": "The actions this guard is allowed to take.", "items": { "enum": [ "block", "report", "replace" ], "type": "string" }, "maxItems": 10, "type": "array", "x-versionadded": "v2.36" }, "allowedStages": { "description": "The stages where the guard can run.", "items": { "enum": [ "prompt", "response" ], "type": "string" }, "maxItems": 16, "type": "array" }, "awsModel": { "description": "AWS model.", "enum": [ "amazon-titan", "anthropic-claude-2", "anthropic-claude-3-haiku", "anthropic-claude-3-sonnet", "anthropic-claude-3-opus", "anthropic-claude-3.5-sonnet-v1", "anthropic-claude-3.5-sonnet-v2", "amazon-nova-lite", "amazon-nova-micro", "amazon-nova-pro" ], "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "awsRegion": { "description": "AWS model region.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "createdAt": { "description": "When the template was created.", "format": "date-time", "type": "string" }, "creatorId": { "description": "The ID of the user who created the guard template.", "type": [ "string", "null" ] }, "creatorName": { "description": "The name of the user who created the guard template.", "maxLength": 1000, "type": "string" }, "description": { "description": "The guard template description.", "maxLength": 4096, "type": "string" }, "errorMessage": { "description": "Error message if the guard configuration is invalid.", "type": [ "string", "null" ] }, "googleModel": { "description": "Google model.", "enum": [ "chat-bison", "google-gemini-1.5-flash", "google-gemini-1.5-pro" ], "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "googleRegion": { "description": "Google model region.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "id": { "description": "The guard template object ID.", "type": "string" }, "intervention": { "description": "The intervention configuration for the guard.", "properties": { "action": { "description": "The action to take if conditions are met.", "enum": [ "block", "report", "replace" ], "type": "string" }, "allowedActions": { "description": "The actions this guard is allowed to take.", "items": { "enum": [ "block", "report", "replace" ], "type": "string" }, "maxItems": 10, "type": "array" }, "conditionLogic": { "default": "any", "description": "The action to take if conditions are met.", "enum": [ "any" ], "type": "string" }, "conditions": { "description": "The list of conditions to trigger intervention.", "items": { "description": "The condition to trigger intervention.", "properties": { "comparand": { "anyOf": [ { "type": "boolean" }, { "type": "number" }, { "type": "string" }, { "items": { "description": "The class name to match.", "maxLength": 128, "type": "string" }, "maxItems": 10, "type": "array" } ], "description": "The condition comparand (basis of comparison)." }, "comparator": { "description": "The condition comparator (operator).", "enum": [ "greaterThan", "lessThan", "equals", "notEquals", "is", "isNot", "matches", "doesNotMatch", "contains", "doesNotContain" ], "type": "string" } }, "required": [ "comparand", "comparator" ], "type": "object", "x-versionadded": "v2.35" }, "maxItems": 1, "type": "array" }, "modifyMessage": { "description": "The message to use if the prompt or response is blocked.", "maxLength": 4096, "type": "string" }, "sendNotification": { "description": "The notification event to create if intervention is triggered.", "type": "boolean" } }, "required": [ "action", "conditions", "modifyMessage" ], "type": "object", "x-versionadded": "v2.35" }, "isAgentic": { "default": true, "description": "True if the guard is suitable for agentic workflows only.", "type": "boolean", "x-versionadded": "v2.37" }, "isValid": { "default": true, "description": "True if the guard is fully configured and valid.", "type": "boolean" }, "labels": { "description": "The list of short strings to associate with the template.", "items": { "description": "A short string to associate with the template.", "maxLength": 255, "type": "string" }, "maxItems": 16, "type": "array", "x-versionadded": "v2.37" }, "llmGatewayModelId": { "description": "The LLM Gateway model ID to use as judge.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "llmType": { "description": "The type of LLM used by this guard.", "enum": [ "openAi", "azureOpenAi", "google", "amazon", "datarobot", "nim", "llmGateway" ], "type": [ "string", "null" ] }, "modelInfo": { "description": "The configuration info for guards using deployed models.", "properties": { "classNames": { "description": "The list of class names for multiclass models.", "items": { "description": "The class name.", "maxLength": 128, "type": "string" }, "maxItems": 100, "type": "array" }, "inputColumnName": { "description": "The input column name.", "maxLength": 255, "type": "string" }, "modelId": { "description": "The ID of the registered model for model guards.", "type": [ "string", "null" ] }, "modelName": { "default": "", "description": "The ID of the registered model for .model guards.", "maxLength": 255, "type": "string" }, "outputColumnName": { "description": "The output column name.", "maxLength": 255, "type": "string" }, "replacementTextColumnName": { "default": "", "description": "The name of the output column with replacement text. Required only if intervention.action is `replace`.", "maxLength": 255, "type": "string" }, "targetType": { "description": "The target type.", "enum": [ "Binary", "Regression", "Multiclass", "TextGeneration" ], "type": "string" } }, "required": [ "inputColumnName", "outputColumnName", "targetType" ], "type": "object", "x-versionadded": "v2.35" }, "name": { "description": "The guard template name.", "maxLength": 255, "type": "string" }, "nemoEvaluatorType": { "description": "The guard template \"NeMo Evaluator\" metric type.", "enum": [ "llm_judge", "context_relevance", "response_groundedness", "topic_adherence", "agent_goal_accuracy", "response_relevancy", "faithfulness" ], "type": [ "string", "null" ], "x-versionadded": "v2.41" }, "nemoInfo": { "description": "The configuration info for NeMo guards.", "properties": { "actions": { "default": "", "description": "The NeMo guardrails actions.", "maxLength": 4096, "type": "string" }, "blockedTerms": { "description": "The NeMo guardrails blocked terms list.", "maxLength": 4096, "type": "string" }, "credentialId": { "description": "The NeMo guardrails credential ID.", "type": [ "string", "null" ] }, "llmPrompts": { "default": "", "description": "The NeMo guardrails prompts.", "maxLength": 4096, "type": "string" }, "mainConfig": { "description": "The overall NeMo configuration YAML.", "maxLength": 4096, "type": "string" }, "railsConfig": { "description": "The NeMo guardrails configuration Colang.", "maxLength": 4096, "type": "string" } }, "required": [ "blockedTerms", "mainConfig", "railsConfig" ], "type": "object", "x-versionadded": "v2.35" }, "nemoLlmJudgeConfig": { "description": "The configuration for the LLM Judge metric.", "properties": { "customMetricDirectionality": { "description": "The custom metric directionality for the LLM judge.", "enum": [ "higherIsBetter", "lowerIsBetter" ], "type": "string", "x-versionadded": "v2.42" }, "scoreParsingRegex": { "description": "The regex to parse the score from the LLM judge response.", "maxLength": 1024, "type": "string" }, "systemPrompt": { "description": "The system prompt for the LLM judge.", "maxLength": 8192, "type": "string" }, "userPrompt": { "description": "The user prompt template for the LLM judge.", "maxLength": 8192, "type": "string" } }, "required": [ "customMetricDirectionality", "scoreParsingRegex", "systemPrompt", "userPrompt" ], "type": "object", "x-versionadded": "v2.42" }, "nemoResponseRelevancyConfig": { "description": "The configuration for the Response Relevancy metric.", "properties": { "embeddingDeploymentId": { "description": "The ID of the embedding model deployment.", "type": "string" } }, "required": [ "embeddingDeploymentId" ], "type": "object", "x-versionadded": "v2.42" }, "nemoTopicAdherenceConfig": { "description": "The configuration for the Topic Adherence metric.", "properties": { "metricMode": { "default": "f1", "description": "The metric calculation mode.", "enum": [ "f1", "recall", "precision" ], "type": "string" }, "referenceTopics": { "description": "The list of reference topics.", "items": { "description": "The reference topic.", "maxLength": 512, "type": "string" }, "maxItems": 100, "minItems": 1, "type": "array" } }, "required": [ "referenceTopics" ], "type": "object", "x-versionadded": "v2.42" }, "ootbType": { "description": "The guard template \"Out of the Box\" metric type.", "enum": [ "token_count", "faithfulness", "rouge_1", "agent_goal_accuracy", "agent_goal_accuracy_with_reference", "cost", "task_adherence", "tool_call_accuracy", "agent_guideline_adherence", "agent_latency", "agent_tokens", "agent_cost", "agentGoalAccuracy", "agentGoalAccuracyWithReference", "taskAdherence", "toolCallAccuracy" ], "type": [ "string", "null" ] }, "openaiApiBase": { "description": "The Azure OpenAI API Base URL.", "maxLength": 255, "type": [ "string", "null" ] }, "openaiApiKey": { "description": "Deprecated; use openai_credential instead.", "maxLength": 255, "type": [ "string", "null" ] }, "openaiDeploymentId": { "description": "The OpenAPI deployment ID.", "maxLength": 255, "type": [ "string", "null" ] }, "orgId": { "description": "Organization ID of the user who created the Guard template.", "type": [ "string", "null" ] }, "playgroundOnly": { "description": "Whether the guard is for playground only, or if it can be used in production and playground.", "type": [ "boolean", "null" ], "x-versionadded": "v2.37" }, "productionOnly": { "description": "Whether the guard is for production only, or if it can be used in production and playground.", "type": [ "boolean", "null" ] }, "type": { "description": "The guard template type.", "enum": [ "guardModel", "nemo", "nemoEvaluator", "ootb", "pii", "userModel" ], "type": "string" } }, "required": [ "allowedStages", "createdAt", "description", "id", "name", "type" ], "type": "object", "x-versionadded": "v2.35" }, "maxItems": 200, "type": "array" }, "next": { "description": "The URL of the next page (if null, there is no next page).", "format": "uri", "type": [ "string", "null" ] }, "previous": { "description": "The URL of the previous page (if null, there is no previous page).", "format": "uri", "type": [ "string", "null" ] }, "totalCount": { "description": "The total number of items across all pages.", "type": "integer" } }, "required": [ "data", "next", "previous", "totalCount" ], "type": "object", "x-versionadded": "v2.35" } ``` ### Properties | Name | Type | Required | Restrictions | Description | | --- | --- | --- | --- | --- | | count | integer | false | | The number of items returned on this page. | | data | [GuardTemplateRetrieveResponse] | true | maxItems: 200 | The list of guard templates. | | next | string,null(uri) | true | | The URL of the next page (if null, there is no next page). | | previous | string,null(uri) | true | | The URL of the previous page (if null, there is no previous page). | | totalCount | integer | true | | The total number of items across all pages. | ## GuardTemplateRetrieveResponse ``` { "properties": { "allowedActions": { "description": "The actions this guard is allowed to take.", "items": { "enum": [ "block", "report", "replace" ], "type": "string" }, "maxItems": 10, "type": "array", "x-versionadded": "v2.36" }, "allowedStages": { "description": "The stages where the guard can run.", "items": { "enum": [ "prompt", "response" ], "type": "string" }, "maxItems": 16, "type": "array" }, "awsModel": { "description": "AWS model.", "enum": [ "amazon-titan", "anthropic-claude-2", "anthropic-claude-3-haiku", "anthropic-claude-3-sonnet", "anthropic-claude-3-opus", "anthropic-claude-3.5-sonnet-v1", "anthropic-claude-3.5-sonnet-v2", "amazon-nova-lite", "amazon-nova-micro", "amazon-nova-pro" ], "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "awsRegion": { "description": "AWS model region.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "createdAt": { "description": "When the template was created.", "format": "date-time", "type": "string" }, "creatorId": { "description": "The ID of the user who created the guard template.", "type": [ "string", "null" ] }, "creatorName": { "description": "The name of the user who created the guard template.", "maxLength": 1000, "type": "string" }, "description": { "description": "The guard template description.", "maxLength": 4096, "type": "string" }, "errorMessage": { "description": "Error message if the guard configuration is invalid.", "type": [ "string", "null" ] }, "googleModel": { "description": "Google model.", "enum": [ "chat-bison", "google-gemini-1.5-flash", "google-gemini-1.5-pro" ], "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "googleRegion": { "description": "Google model region.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.35" }, "id": { "description": "The guard template object ID.", "type": "string" }, "intervention": { "description": "The intervention configuration for the guard.", "properties": { "action": { "description": "The action to take if conditions are met.", "enum": [ "block", "report", "replace" ], "type": "string" }, "allowedActions": { "description": "The actions this guard is allowed to take.", "items": { "enum": [ "block", "report", "replace" ], "type": "string" }, "maxItems": 10, "type": "array" }, "conditionLogic": { "default": "any", "description": "The action to take if conditions are met.", "enum": [ "any" ], "type": "string" }, "conditions": { "description": "The list of conditions to trigger intervention.", "items": { "description": "The condition to trigger intervention.", "properties": { "comparand": { "anyOf": [ { "type": "boolean" }, { "type": "number" }, { "type": "string" }, { "items": { "description": "The class name to match.", "maxLength": 128, "type": "string" }, "maxItems": 10, "type": "array" } ], "description": "The condition comparand (basis of comparison)." }, "comparator": { "description": "The condition comparator (operator).", "enum": [ "greaterThan", "lessThan", "equals", "notEquals", "is", "isNot", "matches", "doesNotMatch", "contains", "doesNotContain" ], "type": "string" } }, "required": [ "comparand", "comparator" ], "type": "object", "x-versionadded": "v2.35" }, "maxItems": 1, "type": "array" }, "modifyMessage": { "description": "The message to use if the prompt or response is blocked.", "maxLength": 4096, "type": "string" }, "sendNotification": { "description": "The notification event to create if intervention is triggered.", "type": "boolean" } }, "required": [ "action", "conditions", "modifyMessage" ], "type": "object", "x-versionadded": "v2.35" }, "isAgentic": { "default": true, "description": "True if the guard is suitable for agentic workflows only.", "type": "boolean", "x-versionadded": "v2.37" }, "isValid": { "default": true, "description": "True if the guard is fully configured and valid.", "type": "boolean" }, "labels": { "description": "The list of short strings to associate with the template.", "items": { "description": "A short string to associate with the template.", "maxLength": 255, "type": "string" }, "maxItems": 16, "type": "array", "x-versionadded": "v2.37" }, "llmGatewayModelId": { "description": "The LLM Gateway model ID to use as judge.", "maxLength": 255, "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "llmType": { "description": "The type of LLM used by this guard.", "enum": [ "openAi", "azureOpenAi", "google", "amazon", "datarobot", "nim", "llmGateway" ], "type": [ "string", "null" ] }, "modelInfo": { "description": "The configuration info for guards using deployed models.", "properties": { "classNames": { "description": "The list of class names for multiclass models.", "items": { "description": "The class name.", "maxLength": 128, "type": "string" }, "maxItems": 100, "type": "array" }, "inputColumnName": { "description": "The input column name.", "maxLength": 255, "type": "string" }, "modelId": { "description": "The ID of the registered model for model guards.", "type": [ "string", "null" ] }, "modelName": { "default": "", "description": "The ID of the registered model for .model guards.", "maxLength": 255, "type": "string" }, "outputColumnName": { "description": "The output column name.", "maxLength": 255, "type": "string" }, "replacementTextColumnName": { "default": "", "description": "The name of the output column with replacement text. Required only if intervention.action is `replace`.", "maxLength": 255, "type": "string" }, "targetType": { "description": "The target type.", "enum": [ "Binary", "Regression", "Multiclass", "TextGeneration" ], "type": "string" } }, "required": [ "inputColumnName", "outputColumnName", "targetType" ], "type": "object", "x-versionadded": "v2.35" }, "name": { "description": "The guard template name.", "maxLength": 255, "type": "string" }, "nemoEvaluatorType": { "description": "The guard template \"NeMo Evaluator\" metric type.", "enum": [ "llm_judge", "context_relevance", "response_groundedness", "topic_adherence", "agent_goal_accuracy", "response_relevancy", "faithfulness" ], "type": [ "string", "null" ], "x-versionadded": "v2.41" }, "nemoInfo": { "description": "The configuration info for NeMo guards.", "properties": { "actions": { "default": "", "description": "The NeMo guardrails actions.", "maxLength": 4096, "type": "string" }, "blockedTerms": { "description": "The NeMo guardrails blocked terms list.", "maxLength": 4096, "type": "string" }, "credentialId": { "description": "The NeMo guardrails credential ID.", "type": [ "string", "null" ] }, "llmPrompts": { "default": "", "description": "The NeMo guardrails prompts.", "maxLength": 4096, "type": "string" }, "mainConfig": { "description": "The overall NeMo configuration YAML.", "maxLength": 4096, "type": "string" }, "railsConfig": { "description": "The NeMo guardrails configuration Colang.", "maxLength": 4096, "type": "string" } }, "required": [ "blockedTerms", "mainConfig", "railsConfig" ], "type": "object", "x-versionadded": "v2.35" }, "nemoLlmJudgeConfig": { "description": "The configuration for the LLM Judge metric.", "properties": { "customMetricDirectionality": { "description": "The custom metric directionality for the LLM judge.", "enum": [ "higherIsBetter", "lowerIsBetter" ], "type": "string", "x-versionadded": "v2.42" }, "scoreParsingRegex": { "description": "The regex to parse the score from the LLM judge response.", "maxLength": 1024, "type": "string" }, "systemPrompt": { "description": "The system prompt for the LLM judge.", "maxLength": 8192, "type": "string" }, "userPrompt": { "description": "The user prompt template for the LLM judge.", "maxLength": 8192, "type": "string" } }, "required": [ "customMetricDirectionality", "scoreParsingRegex", "systemPrompt", "userPrompt" ], "type": "object", "x-versionadded": "v2.42" }, "nemoResponseRelevancyConfig": { "description": "The configuration for the Response Relevancy metric.", "properties": { "embeddingDeploymentId": { "description": "The ID of the embedding model deployment.", "type": "string" } }, "required": [ "embeddingDeploymentId" ], "type": "object", "x-versionadded": "v2.42" }, "nemoTopicAdherenceConfig": { "description": "The configuration for the Topic Adherence metric.", "properties": { "metricMode": { "default": "f1", "description": "The metric calculation mode.", "enum": [ "f1", "recall", "precision" ], "type": "string" }, "referenceTopics": { "description": "The list of reference topics.", "items": { "description": "The reference topic.", "maxLength": 512, "type": "string" }, "maxItems": 100, "minItems": 1, "type": "array" } }, "required": [ "referenceTopics" ], "type": "object", "x-versionadded": "v2.42" }, "ootbType": { "description": "The guard template \"Out of the Box\" metric type.", "enum": [ "token_count", "faithfulness", "rouge_1", "agent_goal_accuracy", "agent_goal_accuracy_with_reference", "cost", "task_adherence", "tool_call_accuracy", "agent_guideline_adherence", "agent_latency", "agent_tokens", "agent_cost", "agentGoalAccuracy", "agentGoalAccuracyWithReference", "taskAdherence", "toolCallAccuracy" ], "type": [ "string", "null" ] }, "openaiApiBase": { "description": "The Azure OpenAI API Base URL.", "maxLength": 255, "type": [ "string", "null" ] }, "openaiApiKey": { "description": "Deprecated; use openai_credential instead.", "maxLength": 255, "type": [ "string", "null" ] }, "openaiDeploymentId": { "description": "The OpenAPI deployment ID.", "maxLength": 255, "type": [ "string", "null" ] }, "orgId": { "description": "Organization ID of the user who created the Guard template.", "type": [ "string", "null" ] }, "playgroundOnly": { "description": "Whether the guard is for playground only, or if it can be used in production and playground.", "type": [ "boolean", "null" ], "x-versionadded": "v2.37" }, "productionOnly": { "description": "Whether the guard is for production only, or if it can be used in production and playground.", "type": [ "boolean", "null" ] }, "type": { "description": "The guard template type.", "enum": [ "guardModel", "nemo", "nemoEvaluator", "ootb", "pii", "userModel" ], "type": "string" } }, "required": [ "allowedStages", "createdAt", "description", "id", "name", "type" ], "type": "object", "x-versionadded": "v2.35" } ``` ### Properties | Name | Type | Required | Restrictions | Description | | --- | --- | --- | --- | --- | | allowedActions | [string] | false | maxItems: 10 | The actions this guard is allowed to take. | | allowedStages | [string] | true | maxItems: 16 | The stages where the guard can run. | | awsModel | string,null | false | | AWS model. | | awsRegion | string,null | false | maxLength: 255 | AWS model region. | | createdAt | string(date-time) | true | | When the template was created. | | creatorId | string,null | false | | The ID of the user who created the guard template. | | creatorName | string | false | maxLength: 1000 | The name of the user who created the guard template. | | description | string | true | maxLength: 4096 | The guard template description. | | errorMessage | string,null | false | | Error message if the guard configuration is invalid. | | googleModel | string,null | false | | Google model. | | googleRegion | string,null | false | maxLength: 255 | Google model region. | | id | string | true | | The guard template object ID. | | intervention | GuardInterventionResponse | false | | The intervention configuration for the guard. | | isAgentic | boolean | false | | True if the guard is suitable for agentic workflows only. | | isValid | boolean | false | | True if the guard is fully configured and valid. | | labels | [string] | false | maxItems: 16 | The list of short strings to associate with the template. | | llmGatewayModelId | string,null | false | maxLength: 255 | The LLM Gateway model ID to use as judge. | | llmType | string,null | false | | The type of LLM used by this guard. | | modelInfo | GuardModelInfoResponse | false | | The configuration info for guards using deployed models. | | name | string | true | maxLength: 255 | The guard template name. | | nemoEvaluatorType | string,null | false | | The guard template "NeMo Evaluator" metric type. | | nemoInfo | GuardNemoInfoResponse | false | | The configuration info for NeMo guards. | | nemoLlmJudgeConfig | GuardNemoLlmJudgeConfigResponse | false | | The configuration for the LLM Judge metric. | | nemoResponseRelevancyConfig | GuardNemoResponseRelevancyConfigResponse | false | | The configuration for the Response Relevancy metric. | | nemoTopicAdherenceConfig | GuardNemoTopicAdherenceConfigResponse | false | | The configuration for the Topic Adherence metric. | | ootbType | string,null | false | | The guard template "Out of the Box" metric type. | | openaiApiBase | string,null | false | maxLength: 255 | The Azure OpenAI API Base URL. | | openaiApiKey | string,null | false | maxLength: 255 | Deprecated; use openai_credential instead. | | openaiDeploymentId | string,null | false | maxLength: 255 | The OpenAPI deployment ID. | | orgId | string,null | false | | Organization ID of the user who created the Guard template. | | playgroundOnly | boolean,null | false | | Whether the guard is for playground only, or if it can be used in production and playground. | | productionOnly | boolean,null | false | | Whether the guard is for production only, or if it can be used in production and playground. | | type | string | true | | The guard template type. | ### Enumerated Values | Property | Value | | --- | --- | | awsModel | [amazon-titan, anthropic-claude-2, anthropic-claude-3-haiku, anthropic-claude-3-sonnet, anthropic-claude-3-opus, anthropic-claude-3.5-sonnet-v1, anthropic-claude-3.5-sonnet-v2, amazon-nova-lite, amazon-nova-micro, amazon-nova-pro] | | googleModel | [chat-bison, google-gemini-1.5-flash, google-gemini-1.5-pro] | | llmType | [openAi, azureOpenAi, google, amazon, datarobot, nim, llmGateway] | | nemoEvaluatorType | [llm_judge, context_relevance, response_groundedness, topic_adherence, agent_goal_accuracy, response_relevancy, faithfulness] | | ootbType | [token_count, faithfulness, rouge_1, agent_goal_accuracy, agent_goal_accuracy_with_reference, cost, task_adherence, tool_call_accuracy, agent_guideline_adherence, agent_latency, agent_tokens, agent_cost, agentGoalAccuracy, agentGoalAccuracyWithReference, taskAdherence, toolCallAccuracy] | | type | [guardModel, nemo, nemoEvaluator, ootb, pii, userModel] | ## GuardToolCallInfo ``` { "description": "How to calculate tool call metrics for this guard.", "properties": { "comparisonMetric": { "description": "How tool calls should be compared for metrics calculations.", "enum": [ "exactMatch", "doNotCompare" ], "type": "string" } }, "required": [ "comparisonMetric" ], "type": "object", "x-versionadded": "v2.37" } ``` How to calculate tool call metrics for this guard. ### Properties | Name | Type | Required | Restrictions | Description | | --- | --- | --- | --- | --- | | comparisonMetric | string | true | | How tool calls should be compared for metrics calculations. | ### Enumerated Values | Property | Value | | --- | --- | | comparisonMetric | [exactMatch, doNotCompare] | ## OverallConfigUpdate ``` { "description": "Overall moderation configuration to push (not specific to one guard)", "properties": { "nemoEvaluatorDeploymentId": { "description": "ID of NeMo Evaluator deployment to use for all NeMo Evaluator guards.", "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "timeoutAction": { "description": "Action to take if timeout occurs", "enum": [ "block", "score" ], "type": "string" }, "timeoutSec": { "description": "Timeout value in seconds for any guard", "minimum": 2, "type": "integer" } }, "required": [ "timeoutAction", "timeoutSec" ], "type": "object", "x-versionadded": "v2.35" } ``` Overall moderation configuration to push (not specific to one guard) ### Properties | Name | Type | Required | Restrictions | Description | | --- | --- | --- | --- | --- | | nemoEvaluatorDeploymentId | string,null | false | | ID of NeMo Evaluator deployment to use for all NeMo Evaluator guards. | | timeoutAction | string | true | | Action to take if timeout occurs | | timeoutSec | integer | true | minimum: 2 | Timeout value in seconds for any guard | ### Enumerated Values | Property | Value | | --- | --- | | timeoutAction | [block, score] | ## OverallModerationConfigurationResponse ``` { "properties": { "entityId": { "description": "The ID of the custom model or playground for this configuration.", "type": "string" }, "entityType": { "description": "The type of the associated entity.", "enum": [ "customModel", "customModelVersion", "playground" ], "type": "string" }, "nemoEvaluatorDeploymentId": { "description": "The ID of the NeMo Evaluator deployment to use for all NeMo Evaluator guards.", "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "timeoutAction": { "description": "The action to take if timeout occurs.", "enum": [ "block", "score" ], "type": "string" }, "timeoutSec": { "description": "The timeout value in seconds for any guard.", "minimum": 2, "type": "integer" }, "updatedAt": { "description": "When the configuration was updated.", "format": "date-time", "type": "string" }, "updaterId": { "description": "The ID of the user who updated the configuration.", "type": [ "string", "null" ] } }, "required": [ "entityId", "entityType", "timeoutAction", "timeoutSec", "updaterId" ], "type": "object", "x-versionadded": "v2.35" } ``` ### Properties | Name | Type | Required | Restrictions | Description | | --- | --- | --- | --- | --- | | entityId | string | true | | The ID of the custom model or playground for this configuration. | | entityType | string | true | | The type of the associated entity. | | nemoEvaluatorDeploymentId | string,null | false | | The ID of the NeMo Evaluator deployment to use for all NeMo Evaluator guards. | | timeoutAction | string | true | | The action to take if timeout occurs. | | timeoutSec | integer | true | minimum: 2 | The timeout value in seconds for any guard. | | updatedAt | string(date-time) | false | | When the configuration was updated. | | updaterId | string,null | true | | The ID of the user who updated the configuration. | ### Enumerated Values | Property | Value | | --- | --- | | entityType | [customModel, customModelVersion, playground] | | timeoutAction | [block, score] | ## OverallModerationConfigurationUpdate ``` { "properties": { "entityId": { "description": "The ID of the custom model or playground for this configuration.", "type": "string" }, "entityType": { "description": "The type of the associated entity.", "enum": [ "customModel", "customModelVersion", "playground" ], "type": "string" }, "nemoEvaluatorDeploymentId": { "description": "The ID of the NeMo Evaluator deployment to use for all NeMo Evaluator guards.", "type": [ "string", "null" ], "x-versionadded": "v2.42" }, "timeoutAction": { "description": "The action to take if timeout occurs.", "enum": [ "block", "score" ], "type": "string" }, "timeoutSec": { "description": "The timeout value in seconds for any guard.", "minimum": 0, "type": "integer" } }, "required": [ "entityId", "entityType", "timeoutAction", "timeoutSec" ], "type": "object", "x-versionadded": "v2.35" } ``` ### Properties | Name | Type | Required | Restrictions | Description | | --- | --- | --- | --- | --- | | entityId | string | true | | The ID of the custom model or playground for this configuration. | | entityType | string | true | | The type of the associated entity. | | nemoEvaluatorDeploymentId | string,null | false | | The ID of the NeMo Evaluator deployment to use for all NeMo Evaluator guards. | | timeoutAction | string | true | | The action to take if timeout occurs. | | timeoutSec | integer | true | minimum: 0 | The timeout value in seconds for any guard. | ### Enumerated Values | Property | Value | | --- | --- | | entityType | [customModel, customModelVersion, playground] | | timeoutAction | [block, score] | ## PredictionEnvironmentInUseResponse ``` { "properties": { "id": { "description": "ID of prediction environment.", "type": "string" }, "name": { "description": "Name of prediction environment.", "type": "string" }, "usedBy": { "description": "Guards using this prediction environment.", "items": { "properties": { "configurationId": { "description": "ID of guard configuration.", "type": "string" }, "deploymentId": { "description": "ID of guard model deployment.", "type": "string" }, "name": { "description": "Name of guard configuration.", "type": "string" } }, "required": [ "configurationId", "deploymentId", "name" ], "type": "object", "x-versionadded": "v2.35" }, "maxItems": 32, "type": "array" } }, "required": [ "id", "name", "usedBy" ], "type": "object", "x-versionadded": "v2.35" } ``` ### Properties | Name | Type | Required | Restrictions | Description | | --- | --- | --- | --- | --- | | id | string | true | | ID of prediction environment. | | name | string | true | | Name of prediction environment. | | usedBy | [DeploymentAndGuardResponse] | true | maxItems: 32 | Guards using this prediction environment. | ## SupportedLlmListResponse ``` { "properties": { "count": { "description": "The number of items returned on this page.", "type": "integer" }, "data": { "description": "The list of supported LLMs for moderation.", "items": { "properties": { "description": { "description": "The description of this LLM.", "maxLength": 1024, "type": "string" }, "id": { "description": "The ID for this LLM.", "maxLength": 1024, "type": "string" }, "llmType": { "description": "The general category of this LLM.", "maxLength": 1024, "type": "string" }, "model": { "description": "The specific model of this LLM.", "maxLength": 1024, "type": "string" }, "name": { "description": "The display name of this LLM.", "maxLength": 1024, "type": "string" }, "provider": { "description": "The provider of access to this LLM.", "maxLength": 1024, "type": "string" }, "vendor": { "description": "The vendor of this LLM.", "maxLength": 1024, "type": "string" } }, "required": [ "description", "id", "llmType", "model", "name", "provider", "vendor" ], "type": "object", "x-versionadded": "v2.35" }, "maxItems": 200, "type": "array" }, "next": { "description": "The URL of the next page (if null, there is no next page).", "format": "uri", "type": [ "string", "null" ] }, "previous": { "description": "The URL of the previous page (if null, there is no previous page).", "format": "uri", "type": [ "string", "null" ] }, "totalCount": { "description": "The total number of items across all pages.", "type": "integer" } }, "required": [ "data", "next", "previous", "totalCount" ], "type": "object", "x-versionadded": "v2.35" } ``` ### Properties | Name | Type | Required | Restrictions | Description | | --- | --- | --- | --- | --- | | count | integer | false | | The number of items returned on this page. | | data | [SupportedLlmResponse] | true | maxItems: 200 | The list of supported LLMs for moderation. | | next | string,null(uri) | true | | The URL of the next page (if null, there is no next page). | | previous | string,null(uri) | true | | The URL of the previous page (if null, there is no previous page). | | totalCount | integer | true | | The total number of items across all pages. | ## SupportedLlmResponse ``` { "properties": { "description": { "description": "The description of this LLM.", "maxLength": 1024, "type": "string" }, "id": { "description": "The ID for this LLM.", "maxLength": 1024, "type": "string" }, "llmType": { "description": "The general category of this LLM.", "maxLength": 1024, "type": "string" }, "model": { "description": "The specific model of this LLM.", "maxLength": 1024, "type": "string" }, "name": { "description": "The display name of this LLM.", "maxLength": 1024, "type": "string" }, "provider": { "description": "The provider of access to this LLM.", "maxLength": 1024, "type": "string" }, "vendor": { "description": "The vendor of this LLM.", "maxLength": 1024, "type": "string" } }, "required": [ "description", "id", "llmType", "model", "name", "provider", "vendor" ], "type": "object", "x-versionadded": "v2.35" } ``` ### Properties | Name | Type | Required | Restrictions | Description | | --- | --- | --- | --- | --- | | description | string | true | maxLength: 1024 | The description of this LLM. | | id | string | true | maxLength: 1024 | The ID for this LLM. | | llmType | string | true | maxLength: 1024 | The general category of this LLM. | | model | string | true | maxLength: 1024 | The specific model of this LLM. | | name | string | true | maxLength: 1024 | The display name of this LLM. | | provider | string | true | maxLength: 1024 | The provider of access to this LLM. | | vendor | string | true | maxLength: 1024 | The vendor of this LLM. |