OpenAIModerationGuardrail

Guardrail that flags or blocks input using OpenAI’s moderation API.

Parameters

Parameter	Type	Default	Description
moderation_model	str	"omni-moderation-latest"	The model to use for moderation.
raise_for_categories	Optional[List[str]]	None	The categories to raise for. When None, any flagged category raises.
api_key	Optional[str]	None	The API key to use for moderation. Defaults to the OPENAI_API_KEY environment variable.

Moderation categories

You can check the current list of moderation categories in OpenAI’s docs.