---
title: Policyai (NeMo Guardrail)
type: guardrail
provider: nvidia
category: safety
source: 'https://github.com/NVIDIA-NeMo/Guardrails'
license: Apache-2.0
tags:
  - guardrail
  - nemo
  - rails
  - colang
dateAdded: '2025-09-01'
featured: false
origin: official
language: en
description: >-
  """ PolicyAI Integration Flows (Colang 2.x) PolicyAI provides content
  moderation and policy enforcement capabilities. For more information, see:
  https
---

```coffee
"""
PolicyAI Integration Flows (Colang 2.x)

PolicyAI provides content moderation and policy enforcement capabilities.
For more information, see: https://musubilabs.ai

Supported features:
- Input moderation: Check user messages against configured policies
- Output moderation: Check bot responses against configured policies
- Tag-based evaluation: Use POLICYAI_TAG_NAME env var to specify policy tag
"""

flow policyai moderation on input
  """Guardrail based on PolicyAI assessment."""
  $result = await CallPolicyaiApiAction(text=$user_message)

  if $result.assessment == "UNSAFE"
    if $system.config.enable_rails_exceptions
      send PolicyAIModerationRailException(message="PolicyAI moderation triggered. Content violated policy: " + $result.category)
    else
      bot refuse to respond
    abort

flow policyai moderation on output
  """Guardrail based on PolicyAI assessment."""
  $result = await CallPolicyaiApiAction(text=$bot_message)

  if $result.assessment == "UNSAFE"
    if $system.config.enable_rails_exceptions
      send PolicyAIModerationRailException(message="PolicyAI moderation triggered. Content violated policy: " + $result.category)
    else
      bot refuse to respond
    abort
```