How RecallIndex Works

Data sources

Two public NHTSA APIs feed every page on this site. The Recalls API provides every formal safety-recall campaign — campaign ID, defect description, affected component, the manufacturer's remedy, and the unit count. The Complaints API provides owner-submitted reports of suspected defects, including the date, the component, a free-text summary, and any reported crash or injury. Both APIs are public and free to query; we mirror nothing from third-party sources.

How we group records

NHTSA records are keyed by VIN-derived make, model, and model year. We group recall campaigns under /make/<manufacturer> pages and individual /vehicle/<model> pages, so a Toyota Camry recall surfaces both on the Toyota make page and the Camry model page. Compare pages (/compare) sit side-by-side comparisons of two vehicles using the same dataset.

What we don't do

We do not adjust NHTSA's underlying numbers. The recall counts you see on RecallIndex match what you would find by querying NHTSA directly. Complaint counts are similarly raw — we do not de-duplicate complaints that may describe the same incident from different perspectives.

NHTSA campaigns are defect-based, not vehicle-based. A vehicle can have several open recalls and also have owner complaints about issues that were never formally recalled. The complaint history on each page is intended to surface those non-recall patterns; severity is self-reported and not verified by NHTSA.

Update cadence

The dataset refreshes daily. New recall campaigns and complaints typically appear on RecallIndex the same day they post at NHTSA.

How to verify

Every recall campaign on this site links back to its NHTSA campaign-ID page. If you ever see a number on RecallIndex that doesn't match what NHTSA shows, NHTSA is the canonical source — please tell us and we'll investigate the discrepancy.