NeMo Guardrails
NVIDIA
Open source guardrails framework for LLM and agent safety with a dedicated Colang DSL.
Curated by Tiberiu ArvaVerified
Strengths
- Purpose-built guardrails DSL
- Conversation flow controls
- NVIDIA ecosystem alignment
Practitioner note
Good fit when teams want code-defined conversational and retrieval guardrails they can run themselves, accepting more engineering ownership than a managed service.
Warning
Governance posture
Reviewed 2026-05-26Self-hosted library; residency fully operator-controlled.
SourceOpen-source Python library deployed by the operator; NVIDIA NIM cloud separate. (self-hosted)
SourceNo built-in audit logging; operators add their own.
Self-hosted library; certification applies to the operator.
Self-hosted library; certification applies to the operator's ISMS.
Self-hosted library; applies to the operator.
OSS library component; deployers carry obligations for systems they build. (role: not-applicable)
Apache-2.0 — permissive, OSI-approved. (low)
SourceChange history
Source-backed and auto-detected events for this tool, newest first.
- ReleaseHigh impact
NeMo Guardrails 0.23.0 expanded IORails with streaming and non-streaming tool-call validation, added a standalone /v1/checks validation endpoint, and improved OpenTelemetry observability for production guardrail pipelines.
Release v0.23.0 · NVIDIA-NeMo/Guardrails - ReleaseNeMo Guardrails adds parallel IORails safety execution
NeMo Guardrails' March 12 release introduced IORails for parallel input/output safety execution, plus OpenAI-compatible server support and standalone async rail validation.
Releases · NVIDIA-NeMo/Guardrails
Explore the category
Compare this tool against the rest of its category and the cloud platform foundation layer.
Return to the category grid to compare governance posture across every tracked tool.
Review Microsoft Foundry, Amazon Bedrock, and Gemini Enterprise Agent Platform as the foundation layer.
Track releases, deprecations, license changes, and other market movements.