Introducing OpenGuardrails: AI Safety Tools for Developers on n8n

Built with n8n
1

Hey n8n devs!

Just launched something new you might like — n8n-nodes-openguardrails — a community node that helps you keep your AI workflows safe and clean with built-in moderation and prompt protection.

What’s OpenGuardrails?

OpenGuardrails is an open-source toolkit for AI safety.
It helps you catch things like:

  • Prompt attacks (jailbreaks, prompt injections, etc.)

  • Unsafe content (violence, hate, adult content, etc.)

  • Data leaks (PII, trade secrets, IP)

Basically — it’s here to make sure your LLM workflows behave nicely.

Why You’ll Care

If you’re using ChatGPT, Claude, or any LLM in n8n, you’ve probably run into one of these:

  • Someone tries to trick your AI with a prompt injection

  • The AI spits out something it shouldn’t

  • Sensitive info leaks in a response

OpenGuardrails checks all that for you — both inputs and outputs — and keeps your workflows safe.

Quick Example

Here’s a simple setup to make your chatbot unbreakable:

1. Webhook (get user input)
   ↓
2. OpenGuardrails - Input Moderation
   ↓
3. IF node → (action = "pass")
   → YES: send to AI
   → NO: return safe response
   ↓
4. AI Response
   ↓
5. OpenGuardrails - Output Moderation
   ↓
6. IF node → (action = "pass")
   → YES: return to user
   → NO: use safe alt

Now your bot checks both ends automatically — easy win.

What You Can Do With It

  • Build safer chatbots (no prompt hacks or unsafe replies)

  • Moderate user-generated content from Slack, Discord, etc.

  • Keep writing assistants or translation bots compliant

  • Protect multi-language conversations (supports 119 languages!)

Setup

1. Install

Settings → Community Nodes → Install
Type: n8n-nodes-openguardrails

Or with Docker:

environment:
  - N8N_COMMUNITY_PACKAGES=n8n-nodes-openguardrails

2. Get an API key

Grab a free one here https://api.openguardrails.com
(Or self-host it — it’s open source!)

Then just add it as a credential in n8n:

Credentials → New → OpenGuardrails API → paste key → Save

Highlights

  • 4 operations: Check Content / Input / Output / Conversation

  • 4 risk levels: none / low / medium / high

  • Configurable actions: continue, stop, or replace

  • Optional tracking: handle repeat offenders

  • Batch support: works with n8n’s multi-item mode

Under the Hood

  • Model: OpenGuardrails-Text-2510 (3.3B params)

  • 119 languages

  • Free plan: 10 req/sec, 10000/month

  • Open source (Apache 2.0)

  • Self-hosting supported

Example Output

{
  "action": "reject",
  "risk_level": "high",
  "categories": ["S9"],
  "suggest_answer": "I can’t help with that request.",
  "was_replaced": true
}

Use it to:

  • route logic (with IF nodes)

  • log risky content

  • warn users

  • block or sanitize text

Self-Hosting

Prefer on-prem or private setup?

git clone https://github.com/openguardrails/openguardrails
docker compose up -d

Then point your n8n node to your local API.

Perfect if you care about privacy or want custom tuning.

Links

Join the Conversation

Would love to hear:

  • How you’re handling AI safety in your workflows

  • What kind of checks you’d like to see

  • Any feature ideas or use cases

Drop a comment — or open an issue on GitHub.

TL;DR:
Free, open-source node for AI safety on n8n.
Blocks prompt attacks, unsafe content, and data leaks.
Easy setup, flexible options, developer-friendly.

Install now: n8n-nodes-openguardrails
Stay safe & keep building cool stuff!

1 Like