# NeMo Guardrails

NVIDIA's open-source toolkit for adding programmable guardrails to LLM-based conversational systems.

NeMo Guardrails is an open-source toolkit from NVIDIA. It adds programmable rails around an LLM application. Input rails check what the user sends, and output rails check what the model returns. Dialog, retrieval and execution rails control the conversation flow, the retrieved content and tool calls. Teams use it to block jailbreak attempts and off-topic or unsafe answers at run time. It needs Python 3.10, 3.11, 3.12 or 3.13. The repository is under the Apache 2.0 licence.

Price: Free, open source (Apache 2.0). Site: https://github.com/NVIDIA-NeMo/Guardrails.

Also on the [web page](https://aigovernancetools.fyi/apps/nemo-guardrails).