Library / SDK
Guardrails & Structure
NeMo Guardrails
1Add 2programmable 3guardrails 4to 5conversational 6LLMs
the six Ws · specification
W1
Who
Maintained by NVIDIA.
W2
What
A toolkit for adding programmable topical, safety, and jailbreak rails to LLM conversations.
W3
Where
A Python library and server embedded in your application stack.
W4
When
Reach for it when constraining LLM conversations to allowed topics and safe responses.
W5
Why
A declarative Colang-based way to enforce dialogue and safety policies.
W6
With
Works with OpenAI, LangChain, and Hugging Face models; uses embeddings for matching.
W7
Watch
Maintained by NVIDIA; open-source (Apache-2.0) and fully self-hostable. Prompts go only to the LLM provider you configure.
alternatives