Why we built a car-checker around public records, not guesses
Updated August 14, 2026 · The methods page: how the grades work and what we leave out.
The problem we kept running into
Every trade-in and every used-car purchase runs on the same asymmetry: the dealer has read the car's file and you haven't. The appraiser has the history report, the auction comps, and a decade of pattern knowledge; you have a book value from a website that never saw your car. The tools that exist either give you a number with no verdict (the value sites) or a verdict with no sources (forums, the salesperson). We wanted the third thing: a verdict you can check.
What public records do differently
Almost everything load-bearing about a used car is already on public record. NHTSA's databases hold the VIN decode, every safety recall, and every owner complaint ever filed — free, government-run, checkable by anyone. NMVTIS, the U.S. DOJ's national title system, holds title brands and odometer readings at title events. When a report says "364 powertrain complaints" or "salvage brand, 2019," you don't have to trust us — the source is named on the line and you can look it up yourself. A fact you can verify beats an opinion you have to believe.
Why a fair-window formula is in a VIN checker at all
Because "what's it worth" and "is this offer fair" are different questions. A dealer works backwards from retail: what they can sell your car for, minus reconditioning, their margin (~10% is our rule of thumb), and selling costs. That formula isn't secret — we publish it line by line — so the report computes the window for your VIN and grades the offer in your hand against it: lowball, fair, or strong. Glass box, not black box: every grade shows its arithmetic.
The choices behind the rest
Fail-closed everywhere. When a data source returns nothing, the report omits the section. Missing data reads as "missing" — never as "clean." Every fact carries its source label, and facts from data we license for internal use never appear in consumer reports at all. Pay-per-report, no subscription — you need this three times a decade, not monthly.
What we left out, and why
No failure-rate claims. Complaint counts are raw volume — popular models accrue more — so we report "NHTSA received N complaints," never "this model is unreliable." No climate inference. We only note salt-state exposure when title records actually show it. No price prediction. We show market ranges and the fair-window math, not a made-up "true value." And the limitation that matters most: VetMyVin is a screen, not an inspection — no database sees a worn clutch. A clean report means "worth the next step," and the next step is a $100–150 pre-purchase inspection. Nothing here is an appraisal or legal advice.
Questions about a specific check? The guides cover the common ones — trade-in math, lowball tests, and what free VIN checks actually catch.