Credit Score API — Integration Guide

How to pull, interpret, and act on credit scores programmatically — for lenders, fintechs, and financial platforms.

Foundations

Why Credit Scores Matter

A credit score is the fastest signal of creditworthiness. In India, scores range from 300 to 900, with 750+ considered good by most lenders. The score is computed by credit bureaus using a statistical model that weighs payment history, credit utilisation, length of credit history, credit mix, and recent enquiries.

For lending businesses, the credit score serves as the first filter in the underwriting funnel. It enables rapid pre-qualification, reduces the cost of full bureau pulls for ineligible applicants, and provides a standardised benchmark across the industry.

A credit score API eliminates the need for manual document collection or third-party intermediaries. Instead of asking a customer to "check their score and share a screenshot," you pull the score programmatically with consent — verified, timestamped, and audit-ready.

This matters at scale. When processing thousands of applications daily, a credit score API becomes the backbone of automated pre-qualification and risk segmentation.

API Response

What a Credit Score API Returns

Credit ScoreA numerical score within the 300–900 range, with the bureau source and computation date.
Score BandQualitative classification — Excellent (750+), Good (700–749), Fair (650–699), Poor (below 650).
Report SummaryTotal accounts, active accounts, credit utilisation, and recent enquiry count in a compact payload.
Risk FlagsBinary indicators for defaults, write-offs, settled accounts, or high enquiry frequency.
Trend DataScore movement over recent months — improving, stable, or deteriorating trajectory.
Consent & AuditConsent timestamp, purpose code, and pull ID for regulatory compliance and internal audit.
Integration

Integrating into Onboarding & Underwriting

1. Capture ConsentCollect explicit user consent at sign-up. Store the consent text, timestamp, and user ID.
2. Pull ScoreCall the credit score API with the user's PAN and consent. Handle timeouts and bureau errors gracefully.
3. Segment UserRoute users into risk tiers based on score bands — auto-approve, manual review, or decline.
4. Store & LogPersist the score, pull timestamp, and bureau source for audit trail and future re-pull comparison.
request.sh
# Pull credit score during onboarding
curl --request POST \
  --url https://api.scofit.app/v1/credit/score \
  --header 'Authorization: Bearer sk_live_...' \
  --header 'Content-Type: application/json' \
  --data '{
    "customer_id": "cust_x1y2z3",
    "pan": "ABCPD1234X",
    "consent": "yes",
    "purpose": "onboarding_screening"
  }'

# → Response
{
  "score": 742,
  "band": "GOOD",
  "bureau": "CRIF",
  "total_accounts": 9,
  "utilisation_pct": 28,
  "last_enquiry": "2025-09-14",
  "risk_flags": [],
  "pull_id": "pull_8a2f4c"
}
Important

Limitations & Consent

✓

Data Lag — Bureau data updates monthly. A score pulled today may not reflect a loan disbursed last week. Factor this into decision logic.

✓

Thin-File Customers — New-to-credit customers may have no score or a low score despite having no defaults. Use alternative data alongside bureau scores.

✓

Consent is Non-Negotiable — RBI requires explicit, purpose-specific consent. Do not pull scores without it. Store consent evidence for at least 8 years.

✓

Multi-Bureau Variance — Scores can differ across bureaus by 50–100 points for the same individual. Use at least two bureaus for critical decisions.

✓

Hard Enquiry Impact — Each hard pull reduces the customer's score slightly. Bundle pulls and avoid repeated pulls on the same applicant.

Registered Office

Plot No. 260, Scheme No. 54, PU4, Behind Malhar Mall, Indore, Madhya Pradesh 452001