ContentFilter

class aws_cdk.aws_bedrock_alpha.ContentFilter(*, input_strength, output_strength, type, input_action=None, input_enabled=None, input_modalities=None, output_action=None, output_enabled=None, output_modalities=None)

Bases: object

(experimental) Interface to declare a content filter.

Parameters:

input_strength (ContentFilterStrength) – (experimental) The strength of the content filter to apply to prompts / user input.
output_strength (ContentFilterStrength) – (experimental) The strength of the content filter to apply to model responses.
type (ContentFilterType) – (experimental) The type of harmful category that the content filter is applied to.
input_action (Optional[GuardrailAction]) – (experimental) The action to take when content is detected in the input. Default: GuardrailAction.BLOCK
input_enabled (Optional[bool]) – (experimental) Whether the content filter is enabled for input. Default: true
input_modalities (Optional[Sequence[ModalityType]]) – (experimental) The input modalities to apply the content filter to. Default: undefined - Applies to text modality
output_action (Optional[GuardrailAction]) – (experimental) The action to take when content is detected in the output. Default: GuardrailAction.BLOCK
output_enabled (Optional[bool]) – (experimental) Whether the content filter is enabled for output. Default: true
output_modalities (Optional[Sequence[ModalityType]]) – (experimental) The output modalities to apply the content filter to. Default: undefined - Applies to text modality

Stability:

experimental

ExampleMetadata:

fixture=default infused

Example:

# Create a guardrail to filter inappropriate content
guardrail = bedrock.Guardrail(self, "bedrockGuardrails",
    guardrail_name="my-BedrockGuardrails",
    description="Legal ethical guardrails."
)

guardrail.add_content_filter(
    type=bedrock.ContentFilterType.SEXUAL,
    input_strength=bedrock.ContentFilterStrength.HIGH,
    output_strength=bedrock.ContentFilterStrength.MEDIUM
)

# Create an agent with the guardrail
agent_with_guardrail = bedrock.Agent(self, "AgentWithGuardrail",
    foundation_model=bedrock.BedrockFoundationModel.ANTHROPIC_CLAUDE_HAIKU_V1_0,
    instruction="You are a helpful and friendly agent that answers questions about literature.",
    guardrail=guardrail
)

Attributes

input_action

(experimental) The action to take when content is detected in the input.

Default:: GuardrailAction.BLOCK
Stability:: experimental

input_enabled

(experimental) Whether the content filter is enabled for input.

Default:: true
Stability:: experimental

input_modalities

(experimental) The input modalities to apply the content filter to.

Default:: undefined - Applies to text modality
Stability:: experimental

input_strength

(experimental) The strength of the content filter to apply to prompts / user input.

Stability:: experimental

output_action

(experimental) The action to take when content is detected in the output.

Default:: GuardrailAction.BLOCK
Stability:: experimental

output_enabled

(experimental) Whether the content filter is enabled for output.

Default:: true
Stability:: experimental

output_modalities

(experimental) The output modalities to apply the content filter to.

Default:: undefined - Applies to text modality
Stability:: experimental

output_strength

(experimental) The strength of the content filter to apply to model responses.

Stability:: experimental

type

(experimental) The type of harmful category that the content filter is applied to.

Stability:: experimental