Roberty Studio
AIAI Agents
Guardrails
GuardrailsConfigure the agent's guardrails to moderate and protect its interactions with users.Read pageHarmful contentBlock or moderate answers containing hate speech, insults, sexual content, violence or misconduct.Read pagePrompt attacksProtect the agent against prompt injection and malicious manipulation.Read pageProfanityBlock improper or offensive language in the agent's answers.Read pageContextual grounding checkMake sure the agent's answers are grounded in the context of the documents you uploaded.Read pageRelevance checkMake sure the agent's answers stay relevant to the conversation.Read pageDenied topicsName the subjects the agent must not discuss under any circumstances.Read pageSensitive informationProtect personal data by having the agent mask or block sensitive information.Read page