POST/v1/check

Analyze content (URL, SMS, email or phone number) and return a verdict with risk score.

🛡️

Multi-Layer Detection Engine

Each analysis goes through 6 specialized detection modules for 99%+ accuracy. Average time: < 300ms.

Request

curl -X POST https://api.trustyn.ai/v1/check \
  -H "X-API-Key: trustyn_live_xxxxxxxx" \
  -H "Content-Type: application/json" \
  -d '{
    "content": "https://bnp-secure.xyz/login",
    "type": "url"
  }'

Request Body

JSON object with the following fields:

contentstringrequired

The content to analyze. Can be a URL, SMS, email, or phone number.

typestring

Content type. If omitted, the API auto-detects.

"url""sms""email""phone"
metadataobject

Additional metadata to enrich analysis (optional).

Content Types

url

URLs and links. Analyzes domain, SSL, typosquatting, homoglyphs, etc.

Ex: https://suspicious.xyz

sms

SMS messages. Detects phishing and scam patterns.

Ex: Your package is waiting...

email

Email content. Analyzes text, links, patterns.

Ex: Urgent: Verify your account...

phone

Phone numbers. Checks community reports.

Ex: +33612345678

Response

json
{
  "id": "chk_7f8a9b2c3d4e5f6g",
  "score": 92,
  "verdict": "scam",
  "confidence": 0.96,
  "scam_type": "phishing_bank",
  "brand_targeted": "BNP Paribas",
  
  "layers": [
    {
      "name": "cache",
      "risk_score": 0,
      "confidence": 0,
      "is_threat": false,
      "details": "Not in cache",
      "execution_time_ms": 2
    },
    {
      "name": "phishing",
      "risk_score": 95,
      "confidence": 0.98,
      "is_threat": true,
      "details": "typosquatting, brand_impersonation, suspicious_tld",
      "execution_time_ms": 15
    },
    {
      "name": "semantic",
      "risk_score": 78,
      "confidence": 0.91,
      "is_threat": true,
      "details": "91% similarity with known banking scam",
      "execution_time_ms": 156
    },
    {
      "name": "fraud",
      "risk_score": 85,
      "confidence": 0.89,
      "is_threat": true,
      "details": "Financial fraud detected (89%)",
      "execution_time_ms": 78
    },
    {
      "name": "community",
      "risk_score": 72,
      "confidence": 0.85,
      "is_threat": true,
      "details": "47 reports, trending up",
      "execution_time_ms": 23
    },
    {
      "name": "deep_ai",
      "risk_score": 0,
      "confidence": 0,
      "is_threat": false,
      "details": "Deep analysis not needed",
      "execution_time_ms": 0
    }
  ],
  
  "concordant_signals": 4,
  "concordance_boost": true,
  
  "reasons": [
    "🚨 Phishing: typosquatting, brand impersonation, suspicious TLD",
    "🔍 91% similarity with known banking scam",
    "⚠️ Financial fraud detected (89%)",
    "👥 47 community reports",
    "📊 4 concordant signals (+20% score)"
  ],
  "advice": "Do not click. Fraudulent site impersonating BNP Paribas. Never share your banking credentials.",
  
  "analysis_time_ms": 274,
  "cached": false,
  "created_at": "2026-01-09T14:30:00.000Z"
}

🛡️ Detection Modules

Each analysis uses up to 6 specialized modules. The "layers" field details the result of each module.

1
cache< 10ms

Trustyn Cache - Known scams database

2
phishing< 50ms

Phishing Detection - Typosquatting, homoglyphs, 200+ protected brands

3
semantic< 200ms

Semantic Analysis - Similarity with known scams

4
fraud< 100ms

Fraud Detection - Multi-category classification

5
community< 100ms

Community Intelligence - Reports and trends

6
deep_ai< 500ms

Deep AI Analysis - Advanced behavioral detection

Response Fields

idstring

Unique analysis identifier

scoreinteger (0-100)

Overall risk score calculated by Trustyn engine

verdict"safe" | "suspect" | "scam"

Final analysis verdict

brand_targetedstring | null

Detected impersonated brand (e.g., "BNP Paribas", "PayPal", "Amazon")

layersarray

Detailed results from each detection module

concordant_signalsinteger

Number of modules that detected a threat (boost if ≥3)

reasonsarray

List of detection reasons

advicestring

Personalized advice for end user

Verdicts

0-30

safe

Safe content. No threat detected by modules.

31-70

suspect

Suspicious content. Caution recommended.

71-100

scam

Confirmed scam. Do not interact.

Try it

Caching

Results are cached for 24 hours. Analyzing the same content again returns cached: true and doesn't count against your quota.

99%+ Accuracy

Thanks to multi-layer analysis and the concordant signals system, Trustyn achieves detection accuracy above 99%. The more modules detect a threat, the higher the score.