Multi-accounting is a common and expensive abuse pattern for SaaS platforms, marketplaces, and gaming sites. One person creates dozens or hundreds of accounts to exploit free trials, promotions, referral programs, or voting systems.

Why It's Hard to Stop

Traditional approaches fail because fraudsters have evolved:

  • Email aliases: Gmail's dot trick (j.ohn vs john) and + aliases create unlimited email addresses
  • Residential proxies: Each account appears to come from a different home IP — clean, no flags
  • Antidetect browsers: Tools like Kameleo and GoLogin create unique device fingerprints for each account
  • VPNs: Cheap VPN subscriptions provide thousands of IP addresses

Each control has blind spots on its own: residential proxies weaken IP rules, aliases weaken email deduplication, challenge solvers weaken CAPTCHAs, and antidetect tools target basic fingerprints.

What Actually Works

The solution is layered detection — combining multiple signals that are hard to fake simultaneously:

  • Device fingerprinting + tampering detection: Detect if the browser fingerprint is spoofed (antidetect browsers leave detectable traces even when spoofing)
  • Network intelligence: Flag VPNs, proxies, and residential proxy networks
  • Account velocity: Compare signup timing and reuse patterns in your own application.
  • Pseudonymous account linking: Pass your own non-identifying accountId to measure distinct accounts associated with a device.

Implementation with Maskbreak

Maskbreak returns network and device signals in one API call; your application can combine them with its own account and behavioural context. When a user signs up:

  1. The Maskbreak SDK collects a security token (invisible to the user)
  2. Your backend sends the token to POST /v1/evaluate
  3. Maskbreak returns whether the connection is suspicious + a persistent visitor ID
  4. You check if that visitor ID has already created an account

Use isSuspicious, reason codes, and linked-account counts to choose between allow, review, or step-up verification. Avoid blocking on one signal without measuring false positives.

Results

Measure the result against a labelled holdout: duplicate-account rate, review rate, false positives, and promotion loss. Layered signals make evasion more expensive, but no detector is perfect and thresholds should be tuned on your own traffic.

Try it free at maskbreak.com — 1,000 requests/hour, no credit card required.