Collateral Score
Combine mileage and region-adjusted market values, depreciation forecasts, title and recall signals, and an explainable collateral risk tier in one response. This is a vehicle assessment: borrower credit checks and loan approval remain in your system.
Collateral Score is available on request. Contact
sales to enable it for your account. A valid API
key and account access are required, including for the free example VIN
3TMKB5FNXRM012539.
Who uses this API
Lenders, credit unions, auto finance platforms and portfolio servicing teams use Collateral Score for origination reviews, collateral monitoring and loan-book re-scoring.
Use cases
- Origination: compare available collateral values and risk signals before funding.
- Loan-term review: inspect projected retail value at maturity alongside your own lending policy.
- Portfolio monitoring: re-score up to 500 vehicles per batch and route flagged results for review.
Score one vehicle
Send POST /v1/vehicle/collateral-score with your API key in the key query parameter and a JSON body. Open the same interactive playground used by the other product guides:
Parameters
| Parameter | Location | Required | Description |
|---|---|---|---|
key | Query | Yes | Your CarsXE API key with Collateral Score enabled. |
vin | JSON body | Yes | A 17-character VIN. |
mileage | JSON body | Yes | Integer from 0 to 1000000, in miles. |
zip | JSON body | Yes | Five-digit US ZIP as a string. Its state sets the regional value adjustment; preserve leading zeros. |
loan_term_months | JSON body | No | Integer from 1 to 96. Adds depreciation_forecast.at_loan_maturity when supplied. |
Score a vehicle
Keep API keys in your server environment. The example VIN is free and never counted; other VINs use your account's Collateral Score allowance.
Read the response
| Field | Meaning |
|---|---|
success, input, timestamp | Success indicator, submitted inputs and response timestamp. |
vin, vehicle | VIN and available year, make, model and trim. |
current_values | Average-condition retail, wholesale and trade-in values, MSRP, publication date and valuation confidence. Mileage and regional adjustments are applied when available. |
depreciation_forecast | Retail-based projections at 12, 24, 36 and 48 months, plus maturity if requested. Includes curve source, key, minimum sample size and refresh date. |
depreciation_velocity | Annual depreciation and percentile versus the vehicle's segment, with segment labels, peer count and interpretation. Lower percentiles indicate slower depreciation. |
risk_flags | Salvage, flood, lemon/buyback and other title brands, odometer anomalies, total-loss reports and open high-severity recalls. total_flags counts known raised flags. |
lending_recommendation | A–D risk_tier, suggested_max_ltv, manual_review_recommended, reasons, rationale and recommendation confidence. |
data_gaps | Signals the assessment could not see. Read these before using the recommendation. |
This illustrative response excerpt shows the shape, with synthetic values; it is not a live appraisal:
{
"success": true,
"vin": "3TMKB5FNXRM012539",
"current_values": {
"retail": 32000,
"wholesale": 28500,
"trade_in": 27000,
"msrp": 35000,
"confidence": "high",
"publish_date": "2026-10-01"
},
"lending_recommendation": {
"risk_tier": "A",
"suggested_max_ltv": 120,
"manual_review_recommended": false,
"reasons": [
"No risk flags raised.",
"Depreciates slower than 75% of its segment."
],
"rationale": "No risk flags raised. Depreciates slower than 75% of its segment. Not assessed: accident history.",
"confidence": "high"
},
"data_gaps": ["accident_history"]
}The interactive API reference contains the complete response schema, including nested forecasts and flags.
Interpret risk and missing data
The current rules return suggested maximum LTV in percentage points: 120 means 120%. Tiers are evaluated from most severe to least severe; reasons identify the rules that fired.
| Tier | Suggested maximum LTV | Example triggers |
|---|---|---|
| A | 120% | No higher-tier rule fired on the available inputs. |
| B | 110% | One risk flag or depreciation percentile from 50 through 75. |
| C | 95% | Other title issues, two or more risk flags, or depreciation percentile above 75. |
| D | 80% | Salvage, flood, lemon/buyback or odometer anomaly. Manual review is recommended. |
These are collateral recommendations, not loan approvals. Apply your own policy and review the reasons, confidence and gaps together.
- Unknown is
null, notfalse. An unavailable signal cannot trigger its rule. A tier of A does not prove that all risks were assessed. - Accident history is not currently available.
risk_flags.accident_countisnullandaccident_historyis listed indata_gaps. Title-brand records do not provide accident history. - Forecasts are bounded by the curve. Projections beyond available data return
null; the API never extrapolates.curve_sourceidentifies a model curve or a segment fallback. The base is always current retail value. - Confidence has separate meanings. Valuation confidence reflects value completeness and mileage adjustment. Recommendation confidence reflects assessed data gaps. Accident-history and depreciation-curve gaps do not alone lower recommendation confidence, so always inspect
data_gapseven when confidence ishigh. - Partial coverage is explicit. Gaps can include
vehicle_history,recalls,depreciation_curveandsegment_percentile. A recalls outage produces a gap instead of failing the whole score.
Batch portfolio scoring
Submit 1–500 vehicles to POST /v1/vehicle/collateral-score/batch. Each vehicle uses the same inputs as a single score. Only one batch can run per account at a time.
curl -X POST 'https://api.carsxe.com/v1/vehicle/collateral-score/batch?key=YOUR_API_KEY' \
-H 'Content-Type: application/json' \
-d '{"vehicles":[{"vin":"3TMKB5FNXRM012539","mileage":42000,"zip":"20878","loan_term_months":48}]}'A successful submission returns 202 with success, batch_id, status and total. Save batch_id and poll using the same account's key:
curl -G 'https://api.carsxe.com/v1/vehicle/collateral-score/batch' \
--data-urlencode 'key=YOUR_API_KEY' \
--data-urlencode 'batch_id=YOUR_BATCH_ID'Poll until status is completed, then inspect succeeded, failed and the per-vehicle results. A completed batch can include failures; check each result rather than assuming all vehicles scored.
Completion webhook
Optionally include webhook_url in the submission JSON. It must be a public HTTPS URL; redirects are not followed. Completion sends event collateral_score.batch.completed with batch_id, status, total, succeeded, failed and timestamp. Retrieve the actual vehicle results with the authenticated batch-status request.
Webhooks are sent once and are not retried. Keep polling as a fallback and use the authenticated status endpoint to verify completion.
Batch billing
The example VIN is free. Usage is checked up front for every other submitted vehicle. Plans that allow overage are billed one Collateral Score call per successfully scored vehicle when the batch completes. Hard-capped and prepaid plans are charged for every non-example vehicle at submission, before the 202 response.
Errors
| Status | Meaning and recovery |
|---|---|
400 | Invalid body, VIN, mileage, ZIP, loan term, batch vehicle list, webhook URL or batch ID. Correct the input. |
401 | Missing or invalid API key. |
404 | Account access is not enabled, no market value exists for the VIN, or a requested batch was not found. Inspect message. |
409 | A batch is already processing for this account. Wait for it to finish before submitting another. |
429 | Usage allowance is insufficient. For a batch, quota is checked against all non-example vehicles. |
500 | Internal server error. |
502 | Market value or title-history provider failure. Retry only when retryable: true is present, respecting Retry-After. |
503 | Collateral Score is temporarily switched off. Nothing is charged; retry later. |
504 | An upstream request timed out. Respect Retry-After before retrying. |
See the Errors guide and the operation-specific API reference for response details.
Assess vehicle collateral by VIN with current values, depreciation forecasts, risk flags, an A–D risk tier, and suggested maximum LTV. Includes single and batch scoring.