OpenGuardrails

Author: openguardrails
Version: 0.0.4 Type: tool

Description

Overview

The OpenGuardrails plugin provides AI moderation and security tools for Dify applications.
It is open-source, free, context-aware, and designed for AI application developers.

OpenGuardrails focuses on:

• Both guardrails LLM and platform are open sourced (Apache 2.0)
• Prompt Injection Attack Detection (based on OWASP TOP 10 LLM Applications)
• Content Safety & Compliance (based on Rich risk categories)
• Sensitive Data Leakage Prevention PII, Business Secrets
• Contextual Semantic Understanding for precise detection
• Runtime Protection with dual deployment modes:
  • API Detection Mode – flexible integration and precise control
  • Security Gateway Mode – WAF-like transparent proxy, zero code change

Core Protection Capabilities

• Prompt Attack Detection
  Jailbreaks, prompt injection, role-playing, rule bypass

• Content Safety Detection
  Context-aware detection for content safety

• Sensitive Data Leakage Prevention PII, Business Secrets。

19 Risk Categories:

• High Risk: Sensitive Politics, Violance Crime and more
• Medium Risk: Hate, Commercial Misconduct and more
• Low Risk: Harassment and more

Open Source Advantages

• Apache 2.0 License
• Free to use
• Data never leaves your environment (local processing)
• Fully supports private deployment

Open Source Repositories:

• Code Repository: github.com/openguardrails/dify-plugin-openguardrails
• Model Repository: huggingface.co/openguardrails

Included Tools

1. check_prompt

OpenGuardrails - Check Prompt

• Detect user input for prompt attacks, jailbreaks, malicious operations and content safety issues based on OWASP TOP 10 LLM Applications and Content safety standards

Input: prompt (user input to the model)

Output:

id:
  type: string
  description: "Unique identifier for the guardrails check"
overall_risk_level:
  type: string
  description: "Overall risk level: no_risk, low_risk, medium_risk, high_risk"
suggest_action:
  type: string
  description: "Suggested action: pass, reject, replace"
suggest_answer:
  type: string
  description: "Suggested alternative answer if action is replace, empty string if not applicable"
categories:
  type: string
  description: "Risk categories, separated by commas"
score:
  type: number
  description: "Detection probability score (0.0-1.0)"

2. check_response_ctx

OpenGuardrails - Check Response Contextual

• Detect AI response data leakage and content safety based on context understanding, including harmful content and compliance risks based on Content safety standards

Input: prompt (user input) + response (AI output)
Output:

id:
  type: string
  description: "Unique identifier for the guardrails check"
overall_risk_level:
  type: string
  description: "Overall risk level: no_risk, low_risk, medium_risk, high_risk"
suggest_action:
  type: string
  description: "Suggested action: pass, reject, replace"
suggest_answer:
  type: string
  description: "Suggested alternative answer if action is replace, empty string if not applicable"
categories:
  type: string
  description: "Risk categories, separated by commas"
score:
  type: number
  description: "Detection probability score (0.0-1.0)"

Configure

To use OpenGuardrails, you need an API Key:

1. Register at the OpenGuardrails Platform: https://OpenGuardrails.com/platform/
2. Log in and go to Application Management → Get your API Key for free
3. Add the API Key to your Dify plugin configuration

Example Usage

Use OpenGuardrails’ OGR Check Prompt and OGR Check Response Contextual tools to protect the input and output of large language models.

Issue Feedback

For more details, workflows, and best practices, please visit:

• OpenGuardrails Official Website
• Code Repository
• Model Repository

If you encounter issues, feel free to open an Issue on GitHub.
For business cooperation, please contact thomas@OpenGuardrails.com