Glossary
NeMo Guardrails
NVIDIA's open framework for programmable safety, topic, and conversation guardrails around LLM applications, using a Colang DSL to define allowed and disallowed conversation flows.
NVIDIA’s guardrails framework, sitting between an application and an LLM. NeMo Guardrails uses Colang, a small DSL for describing canonical user intents, allowed flows between them, and forbidden or escalation flows. Inputs and outputs are checked against the rules before reaching the model or the user, with optional fact-check or moderation steps in the chain.
The pattern complements model-level training. Where Llama Guardsafety-guardrailsMeta's open content-moderation model line, designed to classify prompts and responses against a configurable taxonomy of harms, deployable as an input/output filter. Open full entry classifies content after the fact, NeMo Guardrails enforces conversation-level structure: keep the chatbot on its assigned topic, refuse off-topic requests, escalate ambiguous cases to a human review path. Apache 2.0governanceA permissive open-source license used by most open-weight model releases (Llama from 4 onward partial, Qwen, Mistral, DeepSeek, Falcon), allowing commercial use without acceptable-use restrictions. Open full entry licensed.