Fintech · Fraud · KYC · E-commerce

Flag the address before
the chargeback hits.

GeoClear scores every US address for deliverability, fraud risk, and vacancy — before you ship, before you disburse, before you onboard. One call. No model to train.

See risk docs →
geoclear — address risk response
{
  "deliverability": 0.94,
  "fraud_score": 0.81, ⚠ HIGH
  "vacancy_score": 0.67,
  "address_confidence": 0.96,
  "unit_count": 1,
  "address_type": "residential",
  "zip_type": "Standard",
  "rdi": "Residential",
  "_flags": {
    "high_fraud_risk": true,
    "likely_vacant": true,
    "address_verified": true
  }
}

Three scores. One call.

No training pipeline. No data labeling. No model deployment. GeoClear returns pre-computed risk signals on every US address in under 5ms.

🚨
Fraud Score
Before shipment

Velocity, unit anomaly, and confirmed chargeback history. Returns 0–1. Above 0.6: hold for review. Above 0.8: decline. No training data needed.

0.0 – 0.59 normal · 0.6–0.79 review · 0.8+ decline
📦
Deliverability Score
Address confidence

Will this address receive a package? Based on address quality, placement precision, and query history. Below 0.4: route to pickup or flag for address verification.

0.0 – 0.39 undeliverable · 0.4–0.79 uncertain · 0.8+ confirmed
🏚
Vacancy Signal
Skip tracing

Is the address currently occupied? Low query activity plus address class heuristics. Used for skip tracing, insurance lapse detection, and disbursement risk.

0.0 – 0.39 occupied · 0.4–0.69 uncertain · 0.7+ likely vacant

Where teams deploy it

GeoClear plugs into checkout flows, origination pipelines, and KYC stacks. Here's what fraud and compliance teams build first.

E-commerce
Block ring farms at checkout

High fraud_score combined with high vacancy_score is the signature of reshipping fraud. Flag the address before the label prints — not after the chargeback.

if fraud_score > 0.8 && vacancy_score > 0.65 → hold
Lending / BNPL
Address verification at origination

Deliverability below threshold signals a likely synthetic or forwarded address — a leading indicator of address fraud. Catch it before disbursement, not during collections.

if deliverability < 0.4 → require secondary ID verification
KYC / AML
Confirm the address exists and is occupied

address_confidence plus vacancy_score together prove the customer has a real, active address — a signal required for enhanced due diligence programs.

if address_confidence < 0.7 || vacancy_score > 0.7 → EDD
risk-check.sh
# Single address risk check
curl -s "https://geoclear.io/v1/address?
  street=142+Elm+St
  &city=Chicago
  &state=IL
  &key=YOUR_KEY"


# Fraud-relevant response fields
# ──────────────────────────────
"fraud_score": 0.81
"deliverability": 0.94
"vacancy_score": 0.67
"address_confidence":0.96
"address_type": "residential"
"rdi": "Residential"
"_flags".high_fraud_risk: true

One endpoint.
Instant risk signal.

REST + JSON. No SDK, no configuration, no query language. Send a street address — get back fraud score, deliverability, vacancy, and address confidence in a single response.

All fraud-relevant fields are returned on the same call as address lookup, enrichment, and geocoding. No second request.

  • fraud_score 0–1. Velocity + chargeback history + unit anomaly.
  • deliverability Will a parcel arrive? Address quality + precision score.
  • vacancy_score Is anyone home? Query activity + class heuristics.
  • address_confidence Match quality against 198M+ verified US records.
  • _flags Boolean shorthand: high_fraud_risk, likely_vacant, address_verified.
💡
Availability by plan

Deliverability and address confidence are available on all plans including Free (10,000 lookups/month, no credit card). Fraud score and vacancy signal require Professional $499/mo. See all plans →

Stop the next chargeback.

10,000 address lookups free. No credit card. Fraud score and vacancy available on Professional.

Read the docs →