LLM guardrails
Also known as: guardrails · AI guardrails
LLM guardrails are the controls that surround a model in an app: input/output filters, policy enforcement, least privilege, and human oversight.
- They’re the system safety, not the model’s.
- Enforceable, testable, and layered.
- The difference between a demo and a safe product.
The framing: the model is one component. Guardrails are the architecture around it — and that's where most of the real safety lives.