Experimental
@aws-cdk/aws-bedrock-alpha is an experimental (alpha) package: its API may change in any release, without a major version bump.
ContentFilter Struct
Interface to declare a content filter.
Example
ts
// Create a guardrail to filter inappropriate content
const guardrail = new bedrock.Guardrail(this, 'bedrockGuardrails', {
guardrailName: 'my-BedrockGuardrails',
description: 'Legal ethical guardrails.',
});
guardrail.addContentFilter({
type: bedrock.ContentFilterType.SEXUAL,
inputStrength: bedrock.ContentFilterStrength.HIGH,
outputStrength: bedrock.ContentFilterStrength.MEDIUM,
});
// Create an agent with the guardrail
const agentWithGuardrail = new bedrock.Agent(this, 'AgentWithGuardrail', {
foundationModel: bedrock.BedrockFoundationModel.ANTHROPIC_CLAUDE_HAIKU_V1_0,
instruction: 'You are a helpful and friendly agent that answers questions about literature.',
guardrail: guardrail,
});
Properties
inputActionoptionalGuardrailActioninput_actionoptionalGuardrailActioninputActionoptionalGuardrailActionInputActionoptionalGuardrailActionInputActionoptionalGuardrailAction= GuardrailAction.BLOCK#The action to take when content is detected in the input.
inputEnabledoptionalbooleaninput_enabledoptionalboolinputEnabledoptionalBooleanInputEnabledoptionalboolInputEnabledoptional*bool= true#Whether the content filter is enabled for input.
inputModalitiesoptionalModalityType[]input_modalitiesoptionaltyping.List[ModalityType]inputModalitiesoptionalList<ModalityType>InputModalitiesoptionalModalityType[]InputModalitiesoptional*[]ModalityType#The input modalities to apply the content filter to.
Default: undefined - Applies to text modality
inputStrengthContentFilterStrengthinput_strengthContentFilterStrengthinputStrengthContentFilterStrengthInputStrengthContentFilterStrengthInputStrengthContentFilterStrengthrequired#The strength of the content filter to apply to prompts / user input.
outputActionoptionalGuardrailActionoutput_actionoptionalGuardrailActionoutputActionoptionalGuardrailActionOutputActionoptionalGuardrailActionOutputActionoptionalGuardrailAction= GuardrailAction.BLOCK#The action to take when content is detected in the output.
outputEnabledoptionalbooleanoutput_enabledoptionalbooloutputEnabledoptionalBooleanOutputEnabledoptionalboolOutputEnabledoptional*bool= true#Whether the content filter is enabled for output.
outputModalitiesoptionalModalityType[]output_modalitiesoptionaltyping.List[ModalityType]outputModalitiesoptionalList<ModalityType>OutputModalitiesoptionalModalityType[]OutputModalitiesoptional*[]ModalityType#The output modalities to apply the content filter to.
Default: undefined - Applies to text modality
outputStrengthContentFilterStrengthoutput_strengthContentFilterStrengthoutputStrengthContentFilterStrengthOutputStrengthContentFilterStrengthOutputStrengthContentFilterStrengthrequired#The strength of the content filter to apply to model responses.
typeContentFilterTypetypeContentFilterTypetypeContentFilterTypeTypeContentFilterTypeTypeContentFilterTyperequired#The type of harmful category that the content filter is applied to.