Signup Fraud Evaluation
Real-time risk scoring for trial registrations.
Presets:
Event Payload
POST /api/v1/score
Evaluation Result
Ready
Confidence
99.5%
Threshold Tiers
Low < 25 | Mid 25-65 | High ≥ 65
Record ID
--
Signal Weights
| Feature | Value | Weight |
|---|---|---|
| Submit a payload to view feature weights | ||
24-Hour Velocity Check
1 Event (Normal)
Sliding time-series window
Syndicate Graph Size
Node Size: 1
Disjoint-Set cluster tracking
Device OS & IP Alignment
Verified Match
Class-C subnet & BIN geolocation
Inference Latency
< 15 ms
Vectorized decision engine
API Keys & Quotas
Manage outbound authentication keys for backend integration.
Authentication Required
Sign in or register a verified developer account to generate up to 3 API keys.
Customer Records
Tenant audit directory of scored registration events.
| Customer | IP Address | Device ID | Risk Score | Verdict | Date | |
|---|---|---|---|---|---|---|
| Sign in to view customer evaluations. | ||||||
API Reference & Integration Guide
Step-by-step instructions to integrate real-time fraud risk scoring into your app.
Step-by-Step Integration Walkthrough
Step 1: Get API Key
Sign in with Google/Email & click "+ New Key" under the API Keys tab. Copy your
fk_live_... secret key.
Step 2: Collect Signup Event
Collect customer registration attributes:
name, email, client ip_address, browser/device hash device_id, and card payment_token.
Step 3: Call Scoring API
Send a
POST /api/v1/score request with X-API-Key: fk_live_... and JSON payload. Execution returns in <15ms.
Step 4: Act on 3-Tier Verdict
- ALLOW (<20): Grant instant trial.
- STEP-UP (20-60): Trigger SMS/2FA verification.
- BLOCK (≥60): Reject trial abuse & require upfront card payment.
- STEP-UP (20-60): Trigger SMS/2FA verification.
- BLOCK (≥60): Reject trial abuse & require upfront card payment.
Live Code Snippets
Loading snippet...
Model Benchmarks & Thresholds
Telemetry and classification boundaries.
ROC-AUC
0.941
95% CI: [0.923, 0.958]
Abuse Recall
93.8%
95% CI: [91.2%, 96.0%]
Precision @ Threshold
91.2%
Target ≥ 85% Precision
Inference Latency
< 15ms
P99 Vectorized scoring
Decision Strategy (3-Tier Action Policy)
0.0 - 19.9: ALLOW (GENUINE)
Frictionless access for legitimate, non-reused users.
20.0 - 59.9: STEP-UP REVIEW
Requires step-up 2FA, SMS verification, or CAPTCHA.
60.0 - 100.0: BLOCK ABUSE
Immediate refusal & requires upfront card payment.