New Workflow

K

Duplicate Workflow

Create a custom copy of "Privacy-Minded Router: Enhanced PII Detection Workflow"

Template: Privacy-Minded Router: Enhanced PII Detection Workflow

Workflow JSON

{"id":"NJU1HOhEmf4zUtrg","meta":{"instanceId":"43be9f61b7bb4f51d68445a423b853efd7e8b9e9fd6740b29ef3248f194460da","templateCredsSetupCompleted":true},"name":"Privacy-Minded Router: Enhanced PII Detection Workflow","tags":[],"nodes":[{"id":"enhanced-pii-analyzer","name":"Enhanced PII Pattern Analyzer","type":"n8n-nodes-base.code","onError":"continueRegularOutput","maxTries":3,"position":[-20,300],"parameters":{"jsCode":"// Enhanced PII Detection with ML-like scoring and masking\nconst crypto = require('crypto');\n\n// Enhanced PII patterns with confidence scoring\nconst piiPatterns = {\n  // High confidence patterns\n  ssn: {\n    pattern: /\\b\\d{3}-\\d{2}-\\d{4}\\b/g,\n    confidence: 0.95,\n    severity: 'critical',\n    mask: true\n  },\n  creditCard: {\n    pattern: /\\b(?:\\d{4}[\\s-]?){3}\\d{4}\\b/g,\n    confidence: 0.90,\n    severity: 'critical', \n    mask: true\n  },\n  \n  // Medium confidence patterns\n  email: {\n    pattern: /\\b[A-Za-z0-9._%+-]+@[A-Za-z0-9.-]+\\.[A-Z|a-z]{2,}\\b/g,\n    confidence: 0.85,\n    severity: 'high',\n    mask: false\n  },\n  phone: {\n    pattern: /\\b(?:\\+?1[-.\\s]?)?\$?\\d{3}\$?[-.\\s]?\\d{3}[-.\\s]?\\d{4}\\b/g,\n    confidence: 0.80,\n    severity: 'medium',\n    mask: true\n  },\n  \n  // Context-aware patterns\n  ipAddress: {\n    pattern: /\\b(?:\\d{1,3}\\.){3}\\d{1,3}\\b/g,\n    confidence: 0.70,\n    severity: 'medium',\n    mask: false\n  },\n  zipCode: {\n    pattern: /\\b\\d{5}(?:-\\d{4})?\\b/g,\n    confidence: 0.60,\n    severity: 'low',\n    mask: false\n  },\n  \n  // Enhanced patterns\n  driversLicense: {\n    pattern: /\\b[A-Z]{1,2}\\d{6,8}\\b/g,\n    confidence: 0.75,\n    severity: 'high',\n    mask: true\n  },\n  dateOfBirth: {\n    pattern: /\\b(?:0?[1-9]|1[0-2])[/-](?:0?[1-9]|[12]\\d|3[01])[/-](?:19|20)\\d{2}\\b/g,\n    confidence: 0.70,\n    severity: 'medium',\n    mask: true\n  },\n  \n  // New patterns\n  accountNumber: {\n    pattern: /\\b(?:account|acct)[\\s#:]*\\d{6,12}\\b/gi,\n    confidence: 0.85,\n    severity: 'high',\n    mask: true\n  },\n  medicalId: {\n    pattern: /\\b(?:patient|medical)[\\s#:]*\\d{6,10}\\b/gi,\n    confidence: 0.90,\n    severity: 'critical',\n    mask: true\n  }\n};\n\n// Enhanced context detection\nconst contextPatterns = {\n  financial: /\\b(?:bank|credit|loan|mortgage|investment|portfolio)\\b/gi,\n  medical: /\\b(?:patient|doctor|hospital|diagnosis|medication|treatment)\\b/gi,\n  legal: /\\b(?:attorney|lawyer|case|litigation|settlement|contract)\\b/gi,\n  personal: /\\b(?:family|spouse|children|address|home|personal)\\b/gi\n};\n\n// Masking functions\nfunction maskPII(text, pattern, maskChar = '*') {\n  return text.replace(pattern, (match) => {\n    if (match.length <= 4) return maskChar.repeat(match.length);\n    return match.substring(0, 2) + maskChar.repeat(match.length - 4) + match.substring(match.length - 2);\n  });\n}\n\nfunction generateSessionId() {\n  return crypto.randomBytes(16).toString('hex');\n}\n\nconst results = [];\nconst sessionId = generateSessionId();\n\nfor (const item of items) {\n  const chatInput = item.json.body?.message || item.json.message || item.json.chatInput || JSON.stringify(item.json);\n  const timestamp = new Date().toISOString();\n  \n  if (!chatInput) {\n    results.push({\n      json: {\n        ...item.json,\n        sessionId,\n        timestamp,\n        error: \"No chat input found\",\n        piiDetected: false,\n        detectedPatterns: [],\n        routeToOrchestrator: 2,\n        riskScore: 0,\n        context: 'unknown'\n      }\n    });\n    continue;\n  }\n  \n  const detectedPatterns = [];\n  let maskedContent = chatInput;\n  let hasPII = false;\n  let totalRiskScore = 0;\n  let highestSeverity = 'low';\n  \n  // Detect context\n  let detectedContext = 'general';\n  for (const [contextType, pattern] of Object.entries(contextPatterns)) {\n    if (pattern.test(chatInput)) {\n      detectedContext = contextType;\n      break;\n    }\n  }\n  \n  // Enhanced PII detection with scoring\n  for (const [patternName, config] of Object.entries(piiPatterns)) {\n    const matches = chatInput.match(config.pattern);\n    if (matches && matches.length > 0) {\n      hasPII = true;\n      \n      // Calculate risk score\n      const patternRisk = config.confidence * matches.length;\n      totalRiskScore += patternRisk;\n      \n      // Track highest severity\n      const severityLevels = { low: 1, medium: 2, high: 3, critical: 4 };\n      if (severityLevels[config.severity] > severityLevels[highestSeverity]) {\n        highestSeverity = config.severity;\n      }\n      \n      // Mask content if required\n      if (config.mask) {\n        maskedContent = maskPII(maskedContent, config.pattern);\n      }\n      \n      detectedPatterns.push({\n        type: patternName,\n        count: matches.length,\n        confidence: config.confidence,\n        severity: config.severity,\n        examples: config.mask ? \n          matches.slice(0, 1).map(m => maskPII(m, config.pattern)) : \n          matches.slice(0, 1), // Only 1 example for security\n        masked: config.mask\n      });\n    }\n  }\n  \n  // Determine routing with enhanced logic\n  let routeToOrchestrator;\n  let routingReason;\n  \n  if (!hasPII) {\n    routeToOrchestrator = 2; // Cloud\n    routingReason = \"No PII detected - using cloud model\";\n  } else if (highestSeverity === 'critical' || totalRiskScore > 2.0) {\n    routeToOrchestrator = 1; // Local\n    routingReason = \"Critical PII or high risk score - using local model\";\n  } else if (detectedContext === 'medical' || detectedContext === 'financial') {\n    routeToOrchestrator = 1; // Local\n    routingReason = \"Sensitive context detected - using local model\";\n  } else {\n    routeToOrchestrator = 1; // Local (default for any PII)\n    routingReason = \"PII detected - using local model\";\n  }\n  \n  // Enhanced logging and monitoring\n  const processingMetrics = {\n    processingTime: Date.now(),\n    inputLength: chatInput.length,\n    patternsChecked: Object.keys(piiPatterns).length,\n    patternsDetected: detectedPatterns.length\n  };\n  \n  results.push({\n    json: {\n      // Core data\n      originalMessage: chatInput,\n      maskedMessage: maskedContent,\n      piiDetected: hasPII,\n      detectedPatterns: detectedPatterns,\n      \n      // Enhanced routing\n      routeToOrchestrator: routeToOrchestrator,\n      routingReason: routingReason,\n      \n      // Risk assessment\n      riskScore: Math.round(totalRiskScore * 100) / 100,\n      highestSeverity: highestSeverity,\n      context: detectedContext,\n      \n      // Security & compliance\n      sessionId: sessionId,\n      timestamp: timestamp,\n      processingMetrics: processingMetrics,\n      \n      // Data classification\n      dataClassification: highestSeverity === 'critical' ? 'confidential' : \n                         highestSeverity === 'high' ? 'internal' : 'public',\n      \n      // Compliance flags\n      complianceFlags: {\n        gdpr: detectedContext === 'personal' || hasPII,\n        hipaa: detectedContext === 'medical',\n        pci: detectedPatterns.some(p => p.type === 'creditCard'),\n        sox: detectedContext === 'financial'\n      },\n      \n      // Include safe original data\n      ...item.json\n    }\n  });\n}\n\n// Log processing summary (without PII)\nconsole.log(`Processing Summary:`);\nconsole.log(`- Messages processed: ${results.length}`);\nconsole.log(`- PII detected in: ${results.filter(r => r.json.piiDetected).length} messages`);\nconsole.log(`- High-risk messages: ${results.filter(r => r.json.riskScore > 1.5).length}`);\nconsole.log(`- Session ID: ${sessionId}`);\n\nreturn results;"},"retryOnFail":true,"typeVersion":2},{"id":"enhanced-routing-switch","name":"Enhanced PII Routing Switch","type":"n8n-nodes-base.switch","onError":"continueRegularOutput","position":[480,300],"parameters":{"rules":{"values":[{"outputKey":"Critical PII - Local Only","conditions":{"options":{"version":2,"leftValue":"","caseSensitive":true,"typeValidation":"strict"},"combinator":"or","conditions":[{"id":"critical-pii-condition","operator":{"type":"string","operation":"equals"},"leftValue":"={{ $json.highestSeverity }}","rightValue":"critical"}]},"renameOutput":true},{"outputKey":"PII Detected - Local Processing","conditions":{"options":{"version":2,"leftValue":"","caseSensitive":true,"typeValidation":"strict"},"combinator":"or","conditions":[{"id":"pii-detected-condition","operator":{"type":"boolean","operation":"equals"},"leftValue":"={{ $json.piiDetected }}","rightValue":true},{"id":"high-risk-condition","operator":{"type":"number","operation":"gte"},"leftValue":"={{ $json.riskScore }}","rightValue":1.5}]},"renameOutput":true},{"outputKey":"Clean - Cloud Processing","conditions":{"options":{"version":2,"leftValue":"","caseSensitive":true,"typeValidation":"strict"},"combinator":"and","conditions":[{"id":"no-pii-condition","operator":{"type":"boolean","operation":"equals"},"leftValue":"={{ $json.piiDetected }}","rightValue":false}]},"renameOutput":true}]},"options":{}},"typeVersion":3.2},{"id":"compliance-logger","name":"Compliance & Audit Logger","type":"n8n-nodes-base.code","onError":"continueRegularOutput","position":[720,0],"parameters":{"jsCode":"// Enhanced logging for compliance and monitoring\nconst auditEntries = [];\nconst timestamp = new Date().toISOString();\n\nfor (const item of items) {\n  // Create comprehensive audit log (without actual PII content)\n  const auditEntry = {\n    timestamp: timestamp,\n    sessionId: item.json.sessionId,\n    messageId: require('crypto').randomBytes(8).toString('hex'),\n    \n    // Processing details\n    processingRoute: item.json.routeToOrchestrator === 1 ? 'local' : 'cloud',\n    routingReason: item.json.routingReason,\n    \n    // Risk assessment\n    piiDetected: item.json.piiDetected || false,\n    riskScore: item.json.riskScore || 0,\n    highestSeverity: item.json.highestSeverity || 'none',\n    context: item.json.context || 'general',\n    \n    // Pattern summary (no actual content)\n    patternsSummary: {\n      totalPatterns: item.json.detectedPatterns?.length || 0,\n      patternTypes: item.json.detectedPatterns?.map(p => p.type) || [],\n      severityLevels: item.json.detectedPatterns?.map(p => p.severity) || []\n    },\n    \n    // Compliance flags\n    complianceFlags: item.json.complianceFlags || {},\n    dataClassification: item.json.dataClassification || 'public',\n    \n    // Performance metrics\n    processingMetrics: item.json.processingMetrics || {},\n    \n    // User interaction (safe metadata only)\n    userMetadata: {\n      inputLength: item.json.originalMessage?.length || 0,\n      responseGenerated: true,\n      errorOccurred: false\n    }\n  };\n  \n  auditEntries.push(auditEntry);\n  \n  // Enhanced console logging for monitoring\n  console.log(`=== PRIVACY-AWARE AI AUDIT LOG ===`);\n  console.log(`Timestamp: ${timestamp}`);\n  console.log(`Session: ${auditEntry.sessionId}`);\n  console.log(`Route: ${auditEntry.processingRoute.toUpperCase()}`);\n  console.log(`PII Status: ${auditEntry.piiDetected ? 'DETECTED' : 'CLEAN'}`);\n  console.log(`Risk Score: ${auditEntry.riskScore}`);\n  console.log(`Context: ${auditEntry.context}`);\n  console.log(`Compliance: ${JSON.stringify(auditEntry.complianceFlags)}`);\n  console.log(`Classification: ${auditEntry.dataClassification}`);\n  console.log(`=====================================`);\n}\n\n// Generate summary metrics\nconst summary = {\n  totalMessages: auditEntries.length,\n  piiMessages: auditEntries.filter(e => e.piiDetected).length,\n  localProcessing: auditEntries.filter(e => e.processingRoute === 'local').length,\n  cloudProcessing: auditEntries.filter(e => e.processingRoute === 'cloud').length,\n  highRiskMessages: auditEntries.filter(e => e.riskScore > 1.5).length,\n  complianceBreaches: 0, // Track any compliance issues\n  averageRiskScore: auditEntries.reduce((sum, e) => sum + e.riskScore, 0) / auditEntries.length\n};\n\nconsole.log(`\\n=== SESSION SUMMARY ===`);\nconsole.log(`Total Messages: ${summary.totalMessages}`);\nconsole.log(`PII Detected: ${summary.piiMessages}`);\nconsole.log(`Local Processing: ${summary.localProcessing}`);\nconsole.log(`Cloud Processing: ${summary.cloudProcessing}`);\nconsole.log(`High Risk: ${summary.highRiskMessages}`);\nconsole.log(`Avg Risk Score: ${summary.averageRiskScore.toFixed(2)}`);\nconsole.log(`=======================`);\n\nreturn items.map((item, index) => ({\n  json: {\n    ...item.json,\n    auditEntry: auditEntries[index],\n    sessionSummary: summary,\n    complianceStatus: 'compliant',\n    privacyScore: 100 - (auditEntries[index].riskScore * 10) // Convert risk to privacy score\n  }\n}));"},"typeVersion":2},{"id":"error-handler","name":"Error Handler & Recovery","type":"n8n-nodes-base.code","onError":"continueRegularOutput","position":[1240,20],"parameters":{"jsCode":"// Centralized error handling and recovery\nconst errors = [];\nconst recoveryActions = [];\n\nfor (const item of items) {\n  try {\n    // Check for processing errors\n    if (item.json.error) {\n      errors.push({\n        timestamp: new Date().toISOString(),\n        sessionId: item.json.sessionId,\n        errorType: 'processing_error',\n        errorMessage: item.json.error,\n        severity: 'medium',\n        recoveryAction: 'logged_and_continued'\n      });\n    }\n    \n    // Check for PII detection failures\n    if (!item.json.detectedPatterns && item.json.piiDetected === undefined) {\n      errors.push({\n        timestamp: new Date().toISOString(),\n        sessionId: item.json.sessionId,\n        errorType: 'pii_detection_failure',\n        errorMessage: 'PII detection did not complete properly',\n        severity: 'high',\n        recoveryAction: 'defaulted_to_local_processing'\n      });\n      \n      // Recovery: Default to safe local processing\n      item.json.piiDetected = true;\n      item.json.routeToOrchestrator = 1;\n      item.json.routingReason = 'Error recovery - defaulted to local';\n      \n      recoveryActions.push('defaulted_to_local_processing');\n    }\n    \n    // Check for compliance violations\n    if (item.json.complianceFlags) {\n      const violations = Object.entries(item.json.complianceFlags)\n        .filter(([key, value]) => value === true)\n        .map(([key]) => key);\n      \n      if (violations.length > 0 && item.json.routeToOrchestrator !== 1) {\n        errors.push({\n          timestamp: new Date().toISOString(),\n          sessionId: item.json.sessionId,\n          errorType: 'compliance_violation',\n          errorMessage: `Compliance-sensitive data routed to cloud: ${violations.join(', ')}`,\n          severity: 'critical',\n          recoveryAction: 'force_local_routing'\n        });\n        \n        // Recovery: Force local processing\n        item.json.routeToOrchestrator = 1;\n        item.json.routingReason = 'Compliance violation recovery - forced local';\n        \n        recoveryActions.push('force_local_routing');\n      }\n    }\n    \n  } catch (error) {\n    errors.push({\n      timestamp: new Date().toISOString(),\n      sessionId: item.json.sessionId || 'unknown',\n      errorType: 'unexpected_error',\n      errorMessage: error.message,\n      severity: 'critical',\n      recoveryAction: 'system_fallback'\n    });\n    \n    // System fallback\n    item.json = {\n      ...item.json,\n      error: 'System error - using safe defaults',\n      piiDetected: true,\n      routeToOrchestrator: 1,\n      routingReason: 'System error recovery',\n      riskScore: 10,\n      highestSeverity: 'critical'\n    };\n    \n    recoveryActions.push('system_fallback');\n  }\n}\n\n// Log all errors and recoveries\nif (errors.length > 0) {\n  console.log(`\\n🚨 ERROR REPORT 🚨`);\n  console.log(`Total Errors: ${errors.length}`);\n  console.log(`Recovery Actions: ${recoveryActions.length}`);\n  \n  errors.forEach((error, index) => {\n    console.log(`\\nError ${index + 1}:`);\n    console.log(`  Type: ${error.errorType}`);\n    console.log(`  Severity: ${error.severity}`);\n    console.log(`  Message: ${error.errorMessage}`);\n    console.log(`  Recovery: ${error.recoveryAction}`);\n    console.log(`  Session: ${error.sessionId}`);\n  });\n  \n  console.log(`\\n🔧 RECOVERY SUMMARY:`);\n  const recoveryStats = recoveryActions.reduce((acc, action) => {\n    acc[action] = (acc[action] || 0) + 1;\n    return acc;\n  }, {});\n  console.log(JSON.stringify(recoveryStats, null, 2));\n}\n\nreturn items.map(item => ({\n  json: {\n    ...item.json,\n    errorHandling: {\n      errorsDetected: errors.length,\n      recoveryActionsApplied: recoveryActions.length,\n      systemHealth: errors.length === 0 ? 'healthy' : \n                   errors.filter(e => e.severity === 'critical').length > 0 ? 'critical' : 'degraded'\n    }\n  }\n}));"},"typeVersion":2},{"id":"monitoring-dashboard","name":"Real-time Monitoring Dashboard","type":"n8n-nodes-base.code","onError":"continueRegularOutput","position":[1840,160],"parameters":{"jsCode":"// Real-time monitoring and alerting\nconst alerts = [];\nconst metrics = {\n  timestamp: new Date().toISOString(),\n  performance: {},\n  security: {},\n  compliance: {},\n  system: {}\n};\n\nfor (const item of items) {\n  const processingTime = Date.now() - new Date(item.json.timestamp).getTime();\n  \n  // Performance monitoring\n  metrics.performance = {\n    averageProcessingTime: processingTime,\n    piiDetectionAccuracy: item.json.detectedPatterns ? 100 : 0,\n    routingEfficiency: item.json.routingReason ? 100 : 0,\n    systemLatency: processingTime\n  };\n  \n  // Security monitoring\n  metrics.security = {\n    piiLeakageRisk: item.json.routeToOrchestrator === 2 && item.json.piiDetected ? 100 : 0,\n    dataClassificationAccuracy: item.json.dataClassification ? 100 : 0,\n    privacyScore: item.json.privacyScore || 100,\n    riskScore: item.json.riskScore || 0\n  };\n  \n  // Compliance monitoring\n  const complianceFlags = item.json.complianceFlags || {};\n  metrics.compliance = {\n    gdprCompliance: complianceFlags.gdpr && item.json.routeToOrchestrator === 1 ? 100 : \n                    complianceFlags.gdpr ? 0 : 100,\n    hipaaCompliance: complianceFlags.hipaa && item.json.routeToOrchestrator === 1 ? 100 :\n                     complianceFlags.hipaa ? 0 : 100,\n    pciCompliance: complianceFlags.pci && item.json.routeToOrchestrator === 1 ? 100 :\n                   complianceFlags.pci ? 0 : 100,\n    overallCompliance: Object.values(complianceFlags).every(flag => \n      !flag || item.json.routeToOrchestrator === 1) ? 100 : 0\n  };\n  \n  // System monitoring\n  metrics.system = {\n    errorRate: item.json.errorHandling?.errorsDetected > 0 ? \n               (item.json.errorHandling.errorsDetected / 1) * 100 : 0,\n    recoveryRate: item.json.errorHandling?.recoveryActionsApplied > 0 ? 100 : 0,\n    systemHealth: item.json.errorHandling?.systemHealth || 'healthy',\n    uptime: 100 // Assume 100% for now\n  };\n  \n  // Generate alerts\n  if (metrics.security.piiLeakageRisk > 0) {\n    alerts.push({\n      level: 'CRITICAL',\n      type: 'PII_LEAKAGE_RISK',\n      message: 'PII detected but routed to cloud processing',\n      timestamp: new Date().toISOString(),\n      sessionId: item.json.sessionId,\n      action: 'IMMEDIATE_REVIEW_REQUIRED'\n    });\n  }\n  \n  if (metrics.compliance.overallCompliance < 100) {\n    alerts.push({\n      level: 'HIGH',\n      type: 'COMPLIANCE_VIOLATION',\n      message: 'Compliance requirements not met',\n      timestamp: new Date().toISOString(),\n      sessionId: item.json.sessionId,\n      details: complianceFlags,\n      action: 'AUDIT_TRAIL_REVIEW'\n    });\n  }\n  \n  if (metrics.performance.averageProcessingTime > 5000) {\n    alerts.push({\n      level: 'MEDIUM',\n      type: 'PERFORMANCE_DEGRADATION',\n      message: `Processing time exceeded threshold: ${metrics.performance.averageProcessingTime}ms`,\n      timestamp: new Date().toISOString(),\n      sessionId: item.json.sessionId,\n      action: 'PERFORMANCE_OPTIMIZATION_NEEDED'\n    });\n  }\n}\n\n// Display monitoring dashboard\nconsole.log(`\\n📊 REAL-TIME MONITORING DASHBOARD 📊`);\nconsole.log(`Timestamp: ${metrics.timestamp}`);\nconsole.log(`\\n🚀 PERFORMANCE METRICS:`);\nconsole.log(`  Processing Time: ${metrics.performance.averageProcessingTime}ms`);\nconsole.log(`  PII Detection: ${metrics.performance.piiDetectionAccuracy}%`);\nconsole.log(`  Routing Efficiency: ${metrics.performance.routingEfficiency}%`);\n\nconsole.log(`\\n🔒 SECURITY METRICS:`);\nconsole.log(`  Privacy Score: ${metrics.security.privacyScore}%`);\nconsole.log(`  Risk Score: ${metrics.security.riskScore}`);\nconsole.log(`  PII Leakage Risk: ${metrics.security.piiLeakageRisk}%`);\n\nconsole.log(`\\n⚖️ COMPLIANCE METRICS:`);\nconsole.log(`  GDPR: ${metrics.compliance.gdprCompliance}%`);\nconsole.log(`  HIPAA: ${metrics.compliance.hipaaCompliance}%`);\nconsole.log(`  PCI: ${metrics.compliance.pciCompliance}%`);\nconsole.log(`  Overall: ${metrics.compliance.overallCompliance}%`);\n\nconsole.log(`\\n🛠️ SYSTEM HEALTH:`);\nconsole.log(`  Error Rate: ${metrics.system.errorRate}%`);\nconsole.log(`  Recovery Rate: ${metrics.system.recoveryRate}%`);\nconsole.log(`  Health Status: ${metrics.system.systemHealth}`);\nconsole.log(`  Uptime: ${metrics.system.uptime}%`);\n\nif (alerts.length > 0) {\n  console.log(`\\n🚨 ACTIVE ALERTS (${alerts.length}):`);\n  alerts.forEach((alert, index) => {\n    console.log(`  ${index + 1}. [${alert.level}] ${alert.type}`);\n    console.log(`     ${alert.message}`);\n    console.log(`     Action: ${alert.action}`);\n    console.log(`     Session: ${alert.sessionId}`);\n  });\n} else {\n  console.log(`\\n✅ NO ACTIVE ALERTS - SYSTEM OPERATING NORMALLY`);\n}\n\nconsole.log(`\\n========================================`);\n\nreturn items.map(item => ({\n  json: {\n    ...item.json,\n    monitoring: {\n      metrics: metrics,\n      alerts: alerts,\n      dashboardGenerated: true,\n      lastCheck: new Date().toISOString()\n    }\n  }\n}));"},"typeVersion":2},{"id":"5824055a-a3ba-4f20-a45c-1849cb164a38","name":"When chat message received","type":"@n8n/n8n-nodes-langchain.chatTrigger","position":[-700,60],"webhookId":"dfadeb7b-13c1-4969-9788-74c1a90d75cc","parameters":{"options":{}},"typeVersion":1.1},{"id":"ad6004ea-e216-496f-a292-dbcff70bf3e3","name":"Ollama Chat Model","type":"@n8n/n8n-nodes-langchain.lmChatOllama","position":[2360,520],"parameters":{"model":"llama2:7b","options":{}},"typeVersion":1},{"id":"ab205aa5-76d3-4d39-81de-e379ac1e825c","name":"Simple Memory","type":"@n8n/n8n-nodes-langchain.memoryBufferWindow","position":[1220,700],"parameters":{"contextWindowLength":50},"typeVersion":1.3},{"id":"dbc77760-ba26-430b-b93a-ba1c262a4841","name":"Agent [Edge]","type":"@n8n/n8n-nodes-langchain.agent","position":[1120,500],"parameters":{"options":{}},"typeVersion":2},{"id":"debcae47-c352-4c58-9abc-c1e2dd46c3ea","name":"AI Agent [Private]","type":"@n8n/n8n-nodes-langchain.agent","position":[2360,320],"parameters":{"text":"={{ $json.maskedMessage }}","options":{},"promptType":"define"},"typeVersion":2},{"id":"bbfe5682-a06a-4744-a6f7-df298ea8595c","name":"Sticky Note","type":"n8n-nodes-base.stickyNote","position":[-1360,240],"parameters":{"color":4,"width":980,"height":1480,"content":"## Privacy-Minded Router: Enhanced PII Detection Workflow Concept [START HERE]\n\nBy: Charles Oh, https://www.linkedin.com/in/codetender/\n\n* This living Concept is for DEMONSTRATION purposes only and is designed to be a starter framework for privacy-minded workflows. \n\n* You'll need credentials for your LLM model providers (click on the model on each agent and view the n8n docs for info)\n\n## Understanding Your Privacy-First AI Architecture\n\n🏗️**ARCHITECTURAL OVERVIEW**\n\n**The Core Problem We're Solving**\n\nModern AI systems are powerful but pose privacy risks when handling sensitive data. Organizations need AI capabilities while ensuring:\n\n✅ Sensitive data never leaves secure environments\n✅ Compliance with regulations (GDPR, HIPAA, PCI, SOX)\n✅ Real-time decision making about data sensitivity\n✅ Comprehensive audit trails for regulatory review\n\n**Our Solution/Concept:**\n \nIntelligent Data Classification + Smart Routing\n\nInstead of treating all data the same, we automatically classify each message and intelligently route it to the appropriate processing environment.\n\nThe processing and routing of the user's chat input occurs before any Agent or LLM model is queried.\n\nIf a user's query has detected PII, the request is categorized and logged for Compliance and Auditing prior to processing by Agent. The current setup can easily be routed and customized further for integrations with more robust Enterprise compliance platforms.\n\nNote: This starter framework is still in beta and does not guarantee any compliance with regulations as this is a work in progress.\n\n## Potential Use Case Examples:\n\nFor Healthcare Organizations\n\nHIPAA Compliance: Medical data automatically routed through full audit pipeline\nPatient Privacy: Zero risk of patient information reaching external systems\nAudit Readiness: Complete documentation trail for regulatory reviews\nOperational Efficiency: Clean queries get optimized cloud processing\nCost Optimization: Monitoring overhead only applied to sensitive data\n\nFor Financial Services\n\nPCI/SOX Compliance: Payment and financial data gets complete audit trail\nCustomer Trust: Demonstrable commitment to data protection with real-time monitoring\nRisk Management: Quantified risk assessment with automatic compliance flagging\nPerformance: Clean financial queries leverage advanced cloud AI capabilities\nRegulatory Reporting: Automated compliance monitoring with detailed metrics\n\nFor Legal Firms\n\nAttorney-Client Privilege: Sensitive case information never leaves local systems\nConfidentiality Assurance: Multi-layer protection with comprehensive monitoring\nPractice Efficiency: Research and non-confidential tasks use optimized cloud processing\nEthics Compliance: Meets professional responsibility requirements with audit trails\nDual Benefits: Maximum security for sensitive data, maximum capability for research\n\nFor Enterprise\n\nData Governance: Centralized control with intelligent processing path selection\nGlobal Compliance: GDPR, privacy laws automatically enforced with real-time monitoring\nProductivity: AI capabilities optimized for both security and performance\nCost Optimization: Pay for enterprise monitoring only when processing sensitive data\nProduction Ready: Real credentials, proven architecture, comprehensive monitoring\n"},"typeVersion":1},{"id":"3482b42a-0547-4fd5-bc5f-932891d01a50","name":"Sticky Note1","type":"n8n-nodes-base.stickyNote","position":[-260,480],"parameters":{"width":540,"height":1240,"content":"## Enhanced PII Pattern Analyzer 🧠\n\n**What it does:**\nThe \"brain\" of the system - analyzes every message for sensitive information\n\n**Technology**: \nJavaScript with crypto libraries\n\n**Core Intelligence Features**:\n\nA) 10 PII Pattern Types with Confidence Scoring\n\njavascriptssn: {\n  pattern: /\\b\\d{3}-\\d{2}-\\d{4}\\b/g,     // Regex pattern\n  confidence: 0.95,                       // 95% confidence this is PII\n  severity: 'critical',                   // Highest security level\n  mask: true                              // Hide in logs\n}\n\nB) Context-Aware Detection\n\n- Financial Context: Detects bank, credit, loan, mortgage discussions\n- Medical Context: Identifies patient, doctor, hospital, diagnosis terms\n- Legal Context: Recognizes attorney, case, litigation language\n- Personal Context: Finds family, address, personal information\n\nC) Risk Scoring Algorithm\njavascriptRisk Score = Σ(Pattern Confidence × Pattern Count)\nTotal Risk = SSN(0.95×1) + Email(0.85×1) + Phone(0.80×1) = 2.6\n\nD) Automatic PII Masking\njavascript\"123-45-6789\" → \"12***-**-**89\"  // Preserves structure, hides content\n\nE) Data Classification\n\nConfidential: Critical PII (SSN, Medical ID, Credit Cards)\nInternal: High severity PII (Account numbers, Driver's licenses)\nPublic: Low/no PII content\n\n**Output Example:**\njson{\n  \"originalMessage\": \"My SSN is 123-45-6789\",\n  \"maskedMessage\": \"My SSN is 12***-**-**89\",\n  \"piiDetected\": true,\n  \"riskScore\": 0.95,\n  \"highestSeverity\": \"critical\",\n  \"context\": \"personal\",\n  \"detectedPatterns\": [\n    {\n      \"type\": \"ssn\",\n      \"confidence\": 0.95,\n      \"severity\": \"critical\",\n      \"masked\": true,\n      \"examples\": [\"12***-**-**89\"]\n    }\n  ],\n  \"complianceFlags\": {\n    \"gdpr\": true,\n    \"hipaa\": false,\n    \"pci\": false,\n    \"sox\": false\n  }\n}"},"typeVersion":1},{"id":"a540749b-c283-4f98-9ce4-7d44c4b8ddf8","name":"Sticky Note2","type":"n8n-nodes-base.stickyNote","position":[300,480],"parameters":{"color":5,"width":500,"height":740,"content":"## Enhanced PII Routing Switch 📊\n\n**What it does: **\nMakes intelligent routing decisions based on analysis results\n\n**Technology: **\nn8n Switch Node with advanced rule logic\n\n**3-Tier Decision Logic:**\n\n**Tier 1: Critical PII - Local Only**\njavascriptif (highestSeverity === 'critical' || riskScore > 2.0) {\n  route = \"Critical PII - Local Only\"\n  // SSNs, Medical IDs, Credit Cards, High-risk combinations\n}\n\n**Tier 2: Standard PII - Local Processing**\njavascriptif (piiDetected === true || riskScore >= 1.5 || \n    context === 'medical' || context === 'financial') {\n  route = \"PII Detected - Local Processing\"  \n  // Any PII, sensitive contexts, medium-risk data\n}\n\n**Tier 3: Clean - Cloud Processing**\njavascriptif (piiDetected === false && riskScore < 1.5) {\n  route = \"Clean - Cloud Processing\"\n  // No PII detected, safe for cloud processing\n}\n\n**Why This Approach Works:**\n\nGranular Control: Not just binary PII/no-PII decisions\nContext Awareness: Financial discussions stay local even without explicit PII\nRisk-Based: Combines multiple factors for smarter routing\nCompliance-First: Ensures sensitive contexts never reach cloud"},"typeVersion":1},{"id":"d90f9215-3ac9-42e8-a05c-e3905267461f","name":"Sticky Note3","type":"n8n-nodes-base.stickyNote","position":[520,-1120],"parameters":{"color":2,"width":1640,"height":240,"content":"## Processing Pipeline:\n\nPII Data → Compliance Logger → Error Handler → Monitoring → AI Agent [Private]\n\n**Why This Architecture:**\n\nComplete Audit Trail: Every sensitive message gets full monitoring\nRegulatory Compliance: Meets audit requirements for HIPAA, GDPR, etc.\nError Recovery: Automatic fallbacks if any monitoring component fails\nPerformance Tracking: Full metrics for sensitive data processing"},"typeVersion":1},{"id":"a9769940-0153-414a-a72b-7a3df0a4905d","name":"Sticky Note4","type":"n8n-nodes-base.stickyNote","position":[920,860],"parameters":{"color":6,"width":580,"height":520,"content":"## Key Capabilities:\n\nOpenRouter account configured for cloud models\n\nEnhanced Memory: 50-message context window for better conversations. Easily swapped to any database/Memory integration.\n\nDirect Processing: Bypasses monitoring overhead for optimal performance\n\nFull AI Power: Access to advanced cloud models (GPT-4, Claude, etc.)\n\nClean Data Assurance: Only processes pre-validated non-sensitive content\n\n**Optimized Architecture:**\nClean Data → Agent [Edge] (Direct Path)\n\n**Why This Approach Works:**\n\nPerformance Optimization: No monitoring overhead for clean data\n\nEnhanced Experience: Better conversation flow with larger memory\n\nCost Efficiency: Pay for cloud processing only when safe and beneficial\n\nUser Experience: Maintains full AI capabilities for non-sensitive queries"},"typeVersion":1},{"id":"3664412c-918f-496d-8e6a-5399a9c55753","name":"Sticky Note5","type":"n8n-nodes-base.stickyNote","position":[520,-860],"parameters":{"color":3,"width":500,"height":1020,"content":"## Compliance & Audit Logger 📋\n\nWhat it does: Creates comprehensive, PII-safe audit trails\nTechnology: Advanced JavaScript with cryptographic session tracking\nAudit Trail Components:\nA) Session Tracking\njavascriptsessionId: \"a1b2c3d4e5f6...\",  // Unique session identifier\nmessageId: \"m1n2o3p4...\",       // Per-message tracking\ntimestamp: \"2025-07-11T...\"     // Precise timing\nB) Processing Metadata (No actual PII stored)\njavascript{\n  processingRoute: \"local|cloud\",\n  routingReason: \"Critical PII detected - using local model\",\n  piiDetected: true,\n  riskScore: 2.75,\n  context: \"financial\",\n  dataClassification: \"confidential\"\n}\nC) Compliance Status\njavascriptcomplianceFlags: {\n  gdpr: true,    // EU personal data regulations\n  hipaa: true,   // US healthcare data protection  \n  pci: false,    // Payment card industry standards\n  sox: true      // Financial reporting compliance\n}\nD) Performance Metrics\njavascriptprocessingMetrics: {\n  processingTime: 1247,        // Milliseconds\n  inputLength: 45,             // Character count\n  patternsChecked: 10,         // Number of PII patterns analyzed\n  patternsDetected: 2          // Patterns found\n}\nPrivacy-Safe Logging:\n\nNo PII Content: Only metadata and classifications stored\nMasked Examples: If examples needed, only masked versions\nCompliance Focus: Designed for regulatory audit requirements\nSession Correlation: Track conversations without storing content"},"typeVersion":1},{"id":"e670b576-d808-453d-a33f-d5b2ded2691d","name":"Sticky Note6","type":"n8n-nodes-base.stickyNote","position":[1040,-860],"parameters":{"color":6,"width":540,"height":1020,"content":"## Error Handler & Recovery 🛠️\n \nWhat it does: Ensures system resilience and compliance even during failures\nTechnology: Comprehensive error detection with automatic recovery\nError Detection Categories:\nA) Processing Errors\n\nPII detection failures → Default to local processing (safe fallback)\nModel unavailability → Route to backup processing\nNetwork issues → Retry with exponential backoff\n\nB) Compliance Violations\n\nSensitive data routed to cloud → Force redirect to local\nMissing compliance flags → Apply conservative classification\nData leakage risk → Immediate alert and local-only routing\n\nC) System Errors\n\nUnexpected failures → Safe defaults with full logging\nResource exhaustion → Graceful degradation\nSecurity breaches → Immediate lockdown procedures\n\nRecovery Strategies:\njavascript// Example: Compliance violation recovery\nif (violations.length > 0 && routeToOrchestrator !== 1) {\n  // Force local processing\n  item.json.routeToOrchestrator = 1;\n  item.json.routingReason = 'Compliance violation recovery - forced local';\n  \n  // Log the violation\n  errors.push({\n    type: 'compliance_violation',\n    severity: 'critical',\n    action: 'force_local_routing'\n  });\n}\nWhy This Matters:\n\nFail-Safe Design: System defaults to maximum security\nCompliance Enforcement: Prevents accidental violations\nOperational Resilience: Continues operating despite component failures\nAudit Completeness: All errors and recoveries are logged"},"typeVersion":1},{"id":"caf87df6-41ea-4680-b474-c060abe96ee9","name":"Sticky Note7","type":"n8n-nodes-base.stickyNote","position":[1600,-860],"parameters":{"color":5,"width":560,"height":1180,"content":"## Real-time Monitoring Dashboard 📊\n \nWhat it does: Provides comprehensive system observability and alerting\nTechnology: Advanced metrics collection with intelligent alerting\nMonitoring Categories:\nA) Performance Metrics\njavascriptperformance: {\n  averageProcessingTime: 1247,    // Target: <2000ms\n  piiDetectionAccuracy: 100,      // Target: >95%\n  routingEfficiency: 100,         // Target: 100%\n  systemLatency: 1247            // End-to-end timing\n}\nB) Security Metrics\njavascriptsecurity: {\n  piiLeakageRisk: 0,             // MUST be 0% (critical alert if >0)\n  dataClassificationAccuracy: 100, // Accuracy of classification\n  privacyScore: 95,              // Overall privacy protection\n  riskScore: 2.1                 // Current message risk level\n}\nC) Compliance Metrics\njavascriptcompliance: {\n  gdprCompliance: 100,           // EU data protection\n  hipaaCompliance: 100,          // Healthcare data protection\n  pciCompliance: 100,            // Payment card security\n  overallCompliance: 100         // Must be 100% for audit\n}\nD) System Health\njavascriptsystem: {\n  errorRate: 0,                  // Target: <5%\n  recoveryRate: 100,             // Success rate of error recovery\n  systemHealth: \"healthy\",       // healthy|degraded|critical\n  uptime: 100                    // System availability\n}\nIntelligent Alerting:\nCritical Alerts (Immediate Action Required)\n\nPII Leakage Risk: Sensitive data routed to cloud\nCompliance Violations: Regulatory requirements not met\nSystem Failures: Components not responding\n\nMedium Alerts (Review Required)\n\nPerformance Degradation: Processing time exceeding thresholds\nAccuracy Issues: PII detection false positives/negatives\nResource Constraints: System approaching limits\n\nInformational (Monitoring)\n\nUsage Patterns: Traffic and routing statistics\nPerformance Trends: Historical metric analysis\nCapacity Planning: Resource utilization trends"},"typeVersion":1},{"id":"8b4453f7-4a4a-421a-9449-e4840c409186","name":"OpenRouter Chat Model","type":"@n8n/n8n-nodes-langchain.lmChatOpenRouter","position":[1080,700],"parameters":{"options":{}},"typeVersion":1},{"id":"1241c848-1667-4e74-9341-ecf06a96a4a5","name":"Sticky Note8","type":"n8n-nodes-base.stickyNote","position":[2340,140],"parameters":{"width":400,"content":"## A cleaned LLM Request\n\nIf PII was detected, the original request should now be here. In the previous steps, any detected PII should be masked. That masked version of the message is what we use in this AI Agent (Private)\n"},"typeVersion":1}],"active":false,"pinData":{},"settings":{"executionOrder":"v1","saveManualExecutions":true,"saveExecutionProgress":true,"saveDataErrorExecution":"all","saveDataSuccessExecution":"all"},"versionId":"0bd6474e-4088-40c1-a92c-bdc91c288e92","connections":{"Simple Memory":{"ai_memory":[[{"node":"Agent [Edge]","type":"ai_memory","index":0}]]},"Ollama Chat Model":{"ai_languageModel":[[{"node":"AI Agent [Private]","type":"ai_languageModel","index":0}]]},"OpenRouter Chat Model":{"ai_languageModel":[[{"node":"Agent [Edge]","type":"ai_languageModel","index":0}]]},"Error Handler & Recovery":{"main":[[{"node":"Real-time Monitoring Dashboard","type":"main","index":0}]]},"Compliance & Audit Logger":{"main":[[{"node":"Error Handler & Recovery","type":"main","index":0}]]},"When chat message received":{"main":[[{"node":"Enhanced PII Pattern Analyzer","type":"main","index":0}]]},"Enhanced PII Routing Switch":{"main":[[{"node":"Compliance & Audit Logger","type":"main","index":0}],[{"node":"Compliance & Audit Logger","type":"main","index":0}],[{"node":"Agent [Edge]","type":"main","index":0}]]},"Enhanced PII Pattern Analyzer":{"main":[[{"node":"Enhanced PII Routing Switch","type":"main","index":0}]]},"Real-time Monitoring Dashboard":{"main":[[{"node":"AI Agent [Private]","type":"main","index":0}]]}}}

Coming Soon:

Visual workflow editor
Drag-and-drop node builder
Gene integration panel

Command Palette

Duplicate Workflow