AIGuardrails

Profanity

Block improper or offensive language in the agent's answers.

The Profanity tab enables a filter that detects and blocks improper or offensive language in the agent's answers.

Turn the profanity filter on

Turn on Enable the profanity filter so the agent blocks answers with unsuitable language.

Enabling it makes the example phrases field available.

Add examples of phrases with improper language, to help the filter recognise what counts as offensive in your agent's context. One phrase per line.

The more specific the examples you give, the more accurately the filter matches your domain's vocabulary.