Class MessageModeratorInputGuardrail

java.lang.Object
dev.langchain4j.guardrails.MessageModeratorInputGuardrail
All Implemented Interfaces:
dev.langchain4j.guardrail.Guardrail<dev.langchain4j.guardrail.InputGuardrailRequest, dev.langchain4j.guardrail.InputGuardrailResult>, dev.langchain4j.guardrail.InputGuardrail
public class MessageModeratorInputGuardrail extends Object implements dev.langchain4j.guardrail.InputGuardrail
An InputGuardrail that validates user messages using a ModerationModel to detect potentially harmful, inappropriate, or policy-violating content.

This guardrail checks incoming user messages for content that should be moderated, such as hate speech, violence, self-harm, sexual content, or other categories defined by the moderation model. If the message is flagged by the moderation model, the validation fails with a fatal result, preventing the message from being processed further.

This is useful for ensuring that user inputs comply with content policies before being sent to an LLM or processed by the application.

  • Constructor Summary

    Constructors
    Constructor
    Description
    MessageModeratorInputGuardrail(dev.langchain4j.model.moderation.ModerationModel moderationModel)
    Constructs a new MessageModeratorInputGuardrail with the specified moderation model.

  • Method Summary

    Modifier and Type
    Method
    Description
    dev.langchain4j.guardrail.InputGuardrailResult
    validate(dev.langchain4j.data.message.UserMessage userMessage)
    Validates the provided user message using the configured ModerationModel.

    Methods inherited from class Object

    clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

    Methods inherited from interface dev.langchain4j.guardrail.Guardrail

    name

    Methods inherited from interface dev.langchain4j.guardrail.InputGuardrail

    failure, failure, fatal, fatal, success, successWith, validate

  • Constructor Details

    • MessageModeratorInputGuardrail

      public MessageModeratorInputGuardrail(dev.langchain4j.model.moderation.ModerationModel moderationModel)
      Constructs a new MessageModeratorInputGuardrail with the specified moderation model.
      Parameters:
      moderationModel - the ModerationModel to use for validating user messages. Must not be null.

  • Method Details

    • validate

      public dev.langchain4j.guardrail.InputGuardrailResult validate(dev.langchain4j.data.message.UserMessage userMessage)
      Validates the provided user message using the configured ModerationModel.

      If the moderation model flags the message as inappropriate or policy-violating, this method returns a fatal result with a ModerationException. Otherwise, it returns a successful validation result.

      Specified by:
      validate in interface dev.langchain4j.guardrail.InputGuardrail
      Parameters:
      userMessage - the UserMessage to validate. Must not be null.
      Returns:
      an InputGuardrailResult indicating success if the message passes moderation, or a fatal result with a ModerationException if the message is flagged