For Developers – AIGuard Integration Guide

How It Works

AIGuard intercepts your outgoing prompt, replaces PII with tokens, and safely restores the PII when the LLM responds.

┌─────────────────────────────────────────────────────────────────┐ │ YOUR APPLICATION │ │ │ │ User sends prompt with PII: │ │ "My email is john@example.com and my phone is 555-123-4567" │ │ │ └─────────────────────────────────────────────────────────────────┘ ▼ ┌─────────────────────────────────────────────────────────────────┐ │ AIGUARD BLINDFOLD PROXY │ │ │ │ Step 1: DETECT PII │ │ ✅ Email: john@example.com │ │ ✅ Phone: 555-123-4567 │ │ │ │ Step 2: MASK PII │ │ ✅ [EMAIL_001] instead of john@example.com │ │ ✅ [PHONE_001] instead of 555-123-4567 │ │ │ │ Step 3: FORWARD TO LLM │ │ "My email is [EMAIL_001] and my phone is [PHONE_001]" │ │ │ │ Step 4: RESTORE PII │ │ ✅ [EMAIL_001] → john@example.com │ │ ✅ [PHONE_001] → 555-123-4567 │ │ │ └─────────────────────────────────────────────────────────────────┘ ▼ ┌─────────────────────────────────────────────────────────────────┐ │ YOUR APPLICATION │ │ │ │ Response with PII restored: │ │ "Your email is john@example.com and your phone is 555-123-4567" │ │ │ └─────────────────────────────────────────────────────────────────┘

Prerequisites

Before integrating AIGuard, you need to generate your API key.

Requirement	Description
AIGuard Account	Sign up at app.aiguard.solutions
API Key	Generate from your dashboard (Settings → API Keys)
LLM Provider	OpenAI, Anthropic, or Google Gemini API key

Supported PII Types

AIGuard detects and masks 16 types of sensitive data:

PII Type	Example	Masked As
Email	john@example.com	`[EMAIL_001]`
Phone (US)	555-123-4567	`[PHONE_US_001]`
Phone (Intl)	+33 6 12 34 56 78	`[PHONE_INTERNATIONAL_001]`
SSN	123-45-6789	`[SSN_001]`
Credit Card	4111-1111-1111-1111	`[CREDIT_CARD_001]`
IP Address (v4)	192.168.1.1	`[IP_V4_001]`
API Key	api_key=sk-abc123...	`[API_KEY_001]`
AWS Key	AKIAIOSFODNN7EXAMPLE	`[AWS_KEY_001]`
Private Key	-----BEGIN RSA...	`[PRIVATE_KEY_001]`
Name	John Doe	`[NAME_001]`

Code Examples

Replace your direct LLM calls with AIGuard calls by updating your base URL.

Python / Direct HTTP

python
import requests
import os

AIGUARD_API_KEY = os.getenv("AIGUARD_API_KEY")

def anonymize_prompt(prompt: str, provider: str = "gemini") -> str:
    """Send prompt through AIGuard for PII anonymization"""
    
    response = requests.post(
        "https://api.aiguard.solutions/api/v1/blindfold/anonymize",
        headers={
            "X-API-Key": AIGUARD_API_KEY,
            "Content-Type": "application/json"
        },
        json={
            "prompt": prompt,
            "provider": provider,
            "masking_mode": "full"
        }
    )
    
    if response.status_code == 200:
        data = response.json()
        return data["response"]
    else:
        raise Exception(f"AIGuard error: {response.text}")

Node.js / Express

javascript
const axios = require('axios');

const AIGUARD_API_KEY = process.env.AIGUARD_API_KEY;
const AIGUARD_BASE_URL = 'https://api.aiguard.solutions/api';

async function anonymizePrompt(prompt, provider = 'gemini') {
    try {
        const response = await axios.post(
            `${AIGUARD_BASE_URL}/v1/blindfold/anonymize`,
            {
                prompt: prompt,
                provider: provider,
                masking_mode: 'full'
            },
            {
                headers: {
                    'X-API-Key': AIGUARD_API_KEY,
                    'Content-Type': 'application/json'
                }
            }
        );
        return response.data.response;
    } catch (error) {
        console.error('AIGuard error:', error.message);
        throw error;
    }
}

cURL

bash
curl -X POST https://api.aiguard.solutions/api/v1/blindfold/anonymize \
  -H "X-API-Key: ag_your_api_key_here" \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": "Customer John Doe (john@example.com) requested a refund.",
    "provider": "gemini",
    "masking_mode": "full"
  }'

Troubleshooting

"Invalid API Key"

Verify your API key is correct.
Check if the key is active in your dashboard.
Ensure you're using the right header (X-API-Key).

"Monthly Limit Exceeded"

Check your plan limits in the dashboard.
Upgrade to a higher tier to increase your quota limit.

"Tokens Not Preserved"

Check the LLM response for token mutations.
Use a lower temperature (0.1 recommended) in your LLM request.