Interface: GuardrailProps

March 19, 2026 · View on GitHub

Properties for creating a Guardrail.

Properties

readonly optional blockedInputMessaging?: string

The message to return when the guardrail blocks a prompt.

"Sorry, your query violates our usage policy."

readonly optional blockedOutputsMessaging?: string

The message to return when the guardrail blocks a model response.

"Sorry, I am unable to answer your question because of our usage policy."

readonly optional contentFilters?: ContentFilter[]

The content filters to apply to the guardrail. Note, if one of

readonly optional contextualGroundingFilters?: ContextualGroundingFilter[]

The contextual grounding filters to apply to the guardrail.

readonly optional deniedTopics?: Topic[]

Up to 30 denied topics to block user inputs or model responses associated with the topic.

readonly optional description?: string

The description of the guardrail.

readonly optional kmsKey?: IKey

A custom KMS key to use for encrypting data.

"Your data is encrypted by default with a key that AWS owns and manages for you."

readonly optional managedWordListFilters?: ManagedWordFilter[]

The managed word filters to apply to the guardrail.

readonly name: string

The name of the guardrail.

readonly optional piiFilters?: PIIFilter[]

The PII filters to apply to the guardrail.

readonly optional regexFilters?: RegexFilter[]

The regular expression (regex) filters to apply to the guardrail.

readonly optional wordFilters?: WordFilter[]

The word filters to apply to the guardrail.