Type Alias GuardrailsOptions

Options for applying guardrails to AI responses to enforce specific constraints.

type GuardrailsOptions = {
    custom?: string;
    disablePii?: boolean;
    disableProfanity?: boolean;
    offTopicAnswers?: boolean;
    professionalTone?: boolean;
}

Properties

custom?: string

A custom guardrail instruction, if specified.

disablePii?: boolean

Disables inclusion of personally identifiable information if true; defaults to false.

disableProfanity?: boolean

Disables profanity in the AI response if true; defaults to false.

offTopicAnswers?: boolean

Prevents off-topic answers if true; defaults to false.

professionalTone?: boolean

Enforces a professional tone in the response if true; defaults to false.