Stable ID + Selfie Pattern

Stable ID Pattern and Stable Selfie Pattern are fraud detection checks that identify repeated reuse of the same face across verification attempts within a short time window. Rather than evaluating a single attempt in isolation, this model looks at behavior across multiple attempts to flag coordinated or automated fraud activity.

This check runs automatically on all Identity Verification jobs. There are no developer changes required, no changes to the end user experience, and no additional cost.

How It Works

Each verification attempt is checked against recent activity to see how often the same face has appeared.

• Normal reuse pattern: The face has not appeared an unusual number of times.
• High reuse pattern: The same face has been used repeatedly across recent attempts.

This applies to both the selfie face and the ID face.

Why This Matters

Fraud tactics continue to evolve, and attackers increasingly rely on replayed or automated face media to pass verification. Since fraud operations typically work with a limited pool of faces, repeated reuse is one of the clearest signals of organized fraud.

This model helps detect:

• Fraud rings
• Automated or replay attacks
• Patterns across multiple attempts that a single-attempt check would miss

Expected impact:

• Reduced pass-through of spoof-related fraud
• Reduced manual review volume in selfie-based flows
• A possible slight increase in friction for edge cases

What You'll See

In the Vouched Dashboard

Visual ID Verification jobs now show two additional checks:

• Green checkmark: Normal reuse pattern
• Yellow warning icon: ID or Selfie velocity triggered

In Webhooks and the FindJobs API

If a pattern check is triggered, one of the following errors is returned:

{
  "errorType": "StableIdPatternError",
  "message": "Unstable ID Pattern Detected"
}

{
  "errorType": "StableSelfiePatternError",
  "message": "Unstable Selfie Pattern Detected"
}