GuardRailCreateRequest

pydantic model openapi_client.models.guard_rail_create_request.GuardRailCreateRequest

Request body for creating a new guard rail. Although only GuardRailName and Description are required here, at least one policy field must also be present - DeniedTopics, FiltersConfig, CustomWordFilters, SensitiveInformationFilters, EnableContextualGroundingCheck or AutomatedReasoningPolicies - because Bedrock rejects a guard rail carrying no policy. A body with none returns GE-1034 with the underlying Bedrock message Guardrail must have at least one policy.

Fields:

• automated_reasoning_policies (List[str] | None)

• blocked_message (str | None)

• custom_word_filters (openapi_client.models.custom_word_filters.CustomWordFilters | None)

• denied_topics (List[openapi_client.models.denied_topic.DeniedTopic] | None)

• description (str)

• enable_contextual_grounding_check (bool | None)

• filters_config (List[openapi_client.models.guard_rail_filter_config.GuardRailFilterConfig] | None)

• grounding_threshold (float | int | None)

• guard_rail_name (str)

• is_cross_region_enabled (bool | None)

• policies_confidence_threshold (float | int | None)

• relevance_threshold (float | int | None)

• scope (str | None)

• sensitive_information_filters (openapi_client.models.sensitive_information_filters.SensitiveInformationFilters | None)

• tier (str | None)

field guard_rail_name: StrictStr [Required] (alias 'GuardRailName')

The name of the guard rail.

Constraints:

• strict = True

field description: StrictStr [Required] (alias 'Description')

The description of the guard rail.

Constraints:

• strict = True

field blocked_message: Optional[StrictStr] = None (alias 'BlockedMessage')

The message to display when content is blocked.

field filters_config: Optional[List[GuardRailFilterConfig]] = None (alias 'FiltersConfig')

Configuration for content filters.

field denied_topics: Optional[List[DeniedTopic]] = None (alias 'DeniedTopics')

List of topics that are denied.

field sensitive_information_filters: Optional[SensitiveInformationFilters] = None (alias 'SensitiveInformationFilters')

field custom_word_filters: Optional[CustomWordFilters] = None (alias 'CustomWordFilters')

field enable_contextual_grounding_check: Optional[StrictBool] = None (alias 'EnableContextualGroundingCheck')

Whether to enable contextual grounding check.

field grounding_threshold: Optional[Union[StrictFloat, StrictInt]] = None (alias 'GroundingThreshold')

The threshold for grounding check.

field relevance_threshold: Optional[Union[StrictFloat, StrictInt]] = None (alias 'RelevanceThreshold')

The threshold for relevance check.

field tier: Optional[StrictStr] = None (alias 'Tier')

The tier of the guard rail (CLASSIC or STANDARD). Defaults to CLASSIC if not provided.

field is_cross_region_enabled: Optional[StrictBool] = None (alias 'IsCrossRegionEnabled')

Whether to enable cross-region configuration for the guard rail. If true, adds guardrailProfileIdentifier. STANDARD tier requires this to be true.

field scope: Optional[StrictStr] = None (alias 'Scope')

The scope of the guard rail (global or private). Defaults to private if not provided. Cannot be changed after creation.

field automated_reasoning_policies: Optional[Annotated[List[StrictStr], Field(max_length=2)]] = None (alias 'AutomatedReasoningPolicies')

Optional list of Automated Reasoning policy ARNs (max 2). Requires IsCrossRegionEnabled to be true.

field policies_confidence_threshold: Optional[Union[Annotated[float, Field(le=1, strict=True, ge=0)], Annotated[int, Field(le=1, strict=True, ge=0)]]] = None (alias 'PoliciesConfidenceThreshold')

Confidence threshold for Automated Reasoning policies (0-1). Only valid when AutomatedReasoningPolicies is provided.

to_str()

Returns the string representation of the model using alias

Return type:

str

to_json()

Returns the JSON representation of the model using alias

Return type:

str

classmethod from_json(json_str)

Create an instance of GuardRailCreateRequest from a JSON string

Return type:

Optional[Self]

to_dict()

Return the dictionary representation of the model using alias.

This has the following differences from calling pydantic’s self.model_dump(by_alias=True):

• None is only added to the output dict for nullable fields that were set at model initialization. Other fields with value None are ignored.

Return type:

Dict[str, Any]

classmethod from_dict(obj)

Create an instance of GuardRailCreateRequest from a dict

Return type:

Optional[Self]