GitHub Demo
Production-ready v1.0.0

Security & Fraud Detection Layer
для sharing economy

Device fingerprinting (HMAC-SHA256) + hybrid AI+heuristic trust scoring + threat policy engine. Compliant с 152-ФЗ/GDPR.

Try Demo Read Docs
13 Trust Criteria
-100..+100 Trust Score
3 Threat Levels
152-ФЗ Compliant
trust-score.js 
import { calculateTrustScoreHeuristic } from '@jocmapp/jguard-security/trust-score'

const trustResult = calculateTrustScoreHeuristic(user)
// {
//   trust_score: 72,
//   risk_level: 'low',
//   contribution_score: 86,
//   criteria_breakdown: [
//     { id: 'complaints', score: -24 },
//     { id: 'activity', score: 45 },
//     { id: 'safety_role', score: 120 },
//     // ... 10 more criteria
//   ]
// }

Features

Production-ready security engine с прозрачной формулой и audit trail

Device Fingerprinting

HMAC-SHA256 хэширование device ID. Compliant с 152-ФЗ/GDPR — не хранить raw device data.

  • Deterministic output (одинаковый input → одинаковый output)
  • HMAC-SHA256 (не обратимый)
  • Device matching (shared devices detection)
  • Banned account detection

Hybrid Trust Score

13 критериев с прозрачной формулой + YandexGPT analysis. Прозрачная формула = compliance с 152-ФЗ.

  • 13 взвешенных критериев
  • Гибрид эвристика + AI (YandexGPT)
  • Прозрачная формула (audit trail)
  • Fallback на эвристику если AI недоступен

Device Matching

Detection shared devices и banned accounts на новых устройствах. Уникально в России.

  • Shared device detection
  • Banned account detection
  • Device fingerprinting (HMAC-SHA256)
  • Device matching (informational only)

Threat Policy Engine

3 уровня угрозы с auto-relax. Gradated response — enterprise-level.

  • NORMAL → ELEVATED → LOCKDOWN
  • Auto-relax (1ч → 3ч → 5ч)
  • Rate limiting + additional verification
  • Admin notification

AI Analysis

YandexGPT integration с fallback на эвристику. Suspicious flags + dynamics.

  • YandexGPT analysis
  • Suspicious flags detection
  • Dynamics tracking
  • Fallback на эвристику

Compliance

152-ФЗ/GDPR compliant из коробки. Audit trail для каждого решения.

  • 152-ФЗ (Персональные данные)
  • GDPR (Right to erasure/access)
  • Audit trail
  • User consent required

Pricing

Open source core + SaaS layer для enterprise features

Core

Open source для community

Free MIT License
  • Device fingerprinting (HMAC-SHA256)
  • Trust score algorithm (13 criteria)
  • Threat policy engine
  • TypeScript types
  • Community support
GitHub
Popular

SaaS

Для production

$300 /мес за 100k API calls
  • Всё из Core
  • Device matching (central DB)
  • AI analysis (YandexGPT)
  • Threat policy (auto-relax)
  • Priority support
  • SLA 99.9%
Contact Sales

Enterprise

White-label + compliance

$2000 /мес
  • Всё из SaaS
  • White-label
  • SLA 99.99%
  • Compliance audit
  • On-premise deployment
  • Dedicated support
Contact Sales

API

REST API для trust score, device matching, threat policy

GET /api/v1/trust-score/:userId

Получить trust score пользователя

{
  "trust_score": 72,
  "risk_level": "low",
  "contribution_score": 86,
  "criteria_breakdown": [
    { "id": "complaints", "score": -24 },
    { "id": "activity", "score": 45 }
  ]
}
POST /api/v1/device/fingerprint

Создать device fingerprint

// Request
{
  "userId": 123,
  "androidId": "abc123...",
  "platform": "android"
}

// Response
{
  "key_hash": "a1b2c3d4...",
  "platform": "android",
  "first_seen_at": "2026-10-29T10:00:00Z"
}
GET /api/v1/device/match?userId=123

Проверить совпадение устройств

{
  "is_matched": true,
  "shared_devices": 2,
  "linked_accounts": [
    {
      "id": 456,
      "username": "john_doe",
      "is_blocked": false,
      "shared_devices": 2
    }
  ],
  "overall": {
    "label": "Shared device detected",
    "risk": "medium"
  }
}
POST /api/v1/threat/evaluate

Оценить threat level

// Request
{
  "userId": 123,
  "eventType": "failed_login",
  "metadata": {
    "ip": "192.168.1.1",
    "attempts": 25
  }
}

// Response
{
  "threat_level": "lockdown",
  "reason": "Too many failed logins: 25 in 15m",
  "auto_relax_duration": 3600000,
  "strike_count": 0
}

Demo

Попробуйте JGUARD прямо сейчас

Trust Score Calculator

Результат появится здесь

Contact

Готовы обсудить интеграцию? Свяжитесь с нами

Email

support@jguard.su

Telegram

@jguard_support

GitHub

gitverse.ru/Lexhit/JGUARD-Security

Docs

docs.jguard.su