​
Spam Filtering

ContactsManager SDK provides advanced spam filtering capabilities that automatically clean contact recommendations, ensuring users only see relevant, high-quality contacts for social features and invitations.

​
How It Works

Modern contact books are digital junkyards filled with service numbers, spam contacts, and commercial communications. Our intelligent filtering system uses multiple layers of detection to separate genuine personal contacts from digital noise:

  1. Multi-Layer Architecture: Deterministic rules, pattern recognition, and network intelligence work together
  2. Real-Time Processing: Spam detection happens instantly without impacting performance
  3. Community Intelligence: Leverages crowdsourced data from our open-source spam database
  4. Privacy-Preserving Hashing: Contact PII is hashed locally before any spam checking occurs
  5. Adaptive Learning: Continuously evolves to detect new spam patterns and techniques

​
Key Filtering Capabilities

​
Deterministic Filtering

  • Service Numbers: Automatic removal of 1-800 numbers, customer support lines, and delivery services
  • Vanity Numbers: Detection of formatted commercial numbers like 1-800-FLOWERS or 1-888-NEW-CARS
  • System Emails: Filtering of no-reply addresses, automated notifications, and system-generated contacts
  • Known Spam: Real-time checking against verified spam databases

​
Pattern Recognition

  • Suspicious Domains: Detection of newly registered domains and low-reputation email providers
  • Temporary Emails: Identification of disposable email addresses from services like 10minutemail
  • Bot Communications: Recognition of automated robocall numbers and marketing communications
  • AI-Generated Contacts: Detection of sophisticated fake personas and generated contact information

​
Network Intelligence

  • Quality Scoring: Contacts are ranked based on authenticity indicators and network presence
  • Mutual Connections: Higher scores for contacts who appear in multiple legitimate contact lists
  • Engagement Prediction: Assessment of likelihood for positive response based on contact quality
  • Cross-Platform Correlation: Detection of spam campaigns spanning multiple platforms

​
Open Source Spam Database

ContactsManager maintains a community-driven spam repository at github.com/contactsmanager-io/spam that includes:

  • Verified Spam Numbers: Thousands of confirmed spam phone numbers with confidence scores
  • Email Patterns: No-reply addresses, promotional domains, and suspicious email patterns
  • Domain Intelligence: Reputation data for newly registered and blocked domains
  • Detection Patterns: Regex patterns for automated spam identification

This transparent, community-maintained approach ensures our filtering stays current with emerging threats while maintaining accountability.

​
Privacy Protection

Spam filtering in ContactsManager maintains our privacy-first approach:

  • Hash-Based Matching: Spam database queries use hashed contact identifiers, never raw data
  • No PII Exposure: Original contact information is never exposed and saved on our servers
  • Transparent Operation: Clear indication when contacts are filtered and why

​
Business Impact

​
Improved User Experience

  • Clean Recommendations: Users see only relevant contacts for invitations
  • Reduced Friction: No time wasted scrolling through spam and service numbers
  • Increased Trust: Quality filtering builds confidence in social features
  • Higher Engagement: Better recommendations lead to more successful invitations

​
Growth Acceleration

  • Conversion Rate Improvements: High quality recommendations lead to higher engagement and increase in invitation conversion rates
  • Viral Coefficient Enhancement: Quality recommendations drive more successful referrals
  • Brand Protection: Avoiding spam recommendations protects your app’s reputation
  • Network Effects: Clean contact lists create stronger, more engaged user networks

​
Implementation

Spam filtering is automatically enabled in all ContactsManager SDK recommendation features:

// Automatically filtered recommendations
let recommendations = await ContactsService.shared.fetch(matching: [.peopleToInvite])

// Clean contact discovery
let appUsers = try await ContactsService.shared.getContactsUsingApp(limit: 20)

// Spam-free mutual connections
let potentialConnections = try await ContactsService.shared.getUsersYouMightKnow(limit: 15)

All recommendation APIs include intelligent spam filtering by default, ensuring your users always receive high-quality contact suggestions without any additional configuration.

For more details on implementing spam filtering in your application, visit contactsmanager.io.