Praison AI

Guardrails CLI

CLI commands for guardrails in PraisonAI TypeScript

The praisonai-ts CLI provides the guardrail command for content validation.

Diagram

User

CLI

Guardrails

Quick Start

  1. Simple Usage
    bash
    praisonai-ts guardrail check "Your content here"
  2. With Configuration
    bash
    praisonai-ts guardrail check "Content" --criteria "Must be professional" --json

Guardrails CLI Commands

The praisonai-ts CLI provides the guardrail command for content validation.

Check Content

bash
# Check content against guardrails
praisonai-ts guardrail check "Your content here"

# Check with custom criteria
praisonai-ts guardrail check "Content" --criteria "Must be professional"

# Get JSON output
praisonai-ts guardrail check "Hello world" --json

Example Output:

json
{
  "success": true,
  "data": {
    "status": "passed",
    "score": 0.95,
    "message": "Content passes validation"
  }
}

SDK Usage

For programmatic guardrail usage:

typescript
import { LLMGuardrail, builtinGuardrails } from 'praisonai';

// LLM-based guardrail
const guard = new LLMGuardrail({
  name: 'safety',
  criteria: 'Content must be safe and appropriate'
});
const result = await guard.check('Hello world');

// Built-in guardrails
const maxLength = builtinGuardrails.maxLength(100);
const lengthResult = await maxLength.run('Hello');

For more details, see the Guardrails SDK documentation.

Related

Guardrails

Guardrails overview

 LLM Guardrail

LLM Guardrail overview