What is the Bank Statement Fraud Detection API?

The Fraud Detection API returns structured risk signals and an authenticity score for bank statement documents after they are processed. It helps detect manipulation, synthetic income, balance tampering, and behavioral anomalies—designed for lenders, fintech platforms, and forensic teams.

How do I use fraud detection with the API?

Upload your bank statement via the Bank Statement OCR API (POST /protected/document), create a job with enableBankMode: true, and wait for the job to complete. Once processing finishes, fraud analysis runs automatically. Call GET /protected/document/:documentId/fraudSignals to retrieve risk signals and score.

What risk signals are detected?

Signals can include balance mismatches, repeated or synthetic deposit patterns, PDF metadata inconsistencies, and behavioral anomalies. Each signal has a type, confidence level, and reason code for integration into your underwriting or review workflow.

Bank Statement Fraud Detection API

Detect altered, manipulated, and synthetic bank statements using structured risk signals and behavioral anomaly detection.

Analyze bank statements for manipulation, synthetic income, balance tampering, and behavioral risk signals. Designed for lenders, fintech platforms, and forensic teams.

Start free trial Bank Statement OCR API →

Quick start (fraud detection)

Fraud analysis runs after a bank statement job completes. Upload a document, create a bank statement job, then request fraud signals for that document.

bash

# 1. Upload a bank statement (same as Bank Statement OCR API)
curl -X POST "https://www.docuclipper.com/api/v1/protected/document" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -F "document=@statement.pdf"

# 2. Create a bank statement job, wait for status=Succeeded (or use webhooks)
# 3. Get fraud signals for the document (after job completes, fraud analysis runs)
curl -X GET "https://www.docuclipper.com/api/v1/protected/document/DOCUMENT_ID/fraudSignals" \
  -H "Authorization: Bearer YOUR_API_KEY"

Or, if you already have a processed documentId:

bash

# Get fraud signals for a processed document
curl -X GET "https://www.docuclipper.com/api/v1/protected/document/DOCUMENT_ID/fraudSignals" \
  -H "Authorization: Bearer YOUR_API_KEY"

Response shape

json

// GET /protected/document/:documentId/fraudSignals (200)
{
  "total": 3,
  "signals": [
    {
      "id": 1,
      "documentId": 12345,
      "type": "balance_tampering",
      "confidence": 0.92,
      "reasonCode": "BALANCE_MISMATCH",
      "metadata": { "expected": 5000, "stated": 12000 }
    },
    {
      "type": "synthetic_income",
      "confidence": 0.78,
      "reasonCode": "REPEATED_DEPOSIT_PATTERN"
    }
  ],
  "score": {
    "documentId": 12345,
    "authenticityScore": 67,
    "riskLevel": "medium"
  }
}

Full example (no SDK)

Complete flow: upload → create bank job → poll until Succeeded → GET /protected/document/:documentId/fraudSignals.

javascript

const fs = require('fs');
const BASE = 'https://www.docuclipper.com/api/v1';
const API_KEY = process.env.DOCUCLIPPER_API_KEY;
const auth = (h = {}) => ({ Authorization: `Bearer ${API_KEY}`, ...h });

const form = new FormData();
form.append('document', new Blob([fs.readFileSync('statement.pdf')]), 'statement.pdf');
const up = await fetch(`${BASE}/protected/document`, { method: 'POST', headers: auth(), body: form });
const docId = (await up.json()).document.id;

const job = await fetch(`${BASE}/protected/job`, {
  method: 'POST', headers: auth({ 'Content-Type': 'application/json' }),
  body: JSON.stringify({ jobName: '', documents: [docId], enableBankMode: true })
});
const jobId = (await job.json()).id;

let status;
while ((status = (await fetch(`${BASE}/protected/job/${jobId}`, { headers: auth() }).then(r => r.json())).status) !== 'Succeeded') {
  if (status === 'Failed') throw new Error('Job failed');
  await new Promise(r => setTimeout(r, 2000));
}

const fraud = await fetch(`${BASE}/protected/document/${docId}/fraudSignals`, { headers: auth() });
const { signals, score } = await fraud.json();
console.log('Score:', score?.authenticityScore, 'Signals:', signals?.length);

Query parameters (fraudSignals)

type – Filter signals by type (e.g. balance_tampering, synthetic_income).
minConfidence / maxConfidence – Filter by confidence (0–1).
limit / offset – Limit results (default limit 100).

Use cases

Loan underwriting and income verification
Fintech KYC and document authenticity checks
Forensic accounting and dispute resolution
Mortgage and tenant screening

Bank Statement OCR API →Webhooks →