Safety rules that limit what an AI can read, say, or do.
AI Guardrails are like a bouncer at a school dance. You can request songs. You cannot drive the DJ booth home.
They block unsafe prompts, replies, and actions. Chatbots and AI Agents use them.
Alignment
AI Guardrails turn Alignment goals into rules an AI must follow.
AI Content Filter
An AI Content Filter can block unsafe input or output.
Agent Permissions
Agent Permissions limit the tools and data an AI Agent can use.