How We Research

RecallScanner exists to turn NHTSA's vehicle safety data into something a normal person can actually use without clicking through twelve different government screens. This page explains where our data comes from, what we do with it, what we add on top, and what we do when something's wrong.

1. Where the data comes from

Every recall campaign, every VIN decode, and every owner complaint displayed on RecallScanner ultimately traces back to a public API operated by the National Highway Traffic Safety Administration (NHTSA). Specifically, we use:

• NHTSA Recalls API — for the recall campaign records you see on brand and model pages, including campaign numbers, affected model years, component categories, consequence language, and remedy instructions.
• NHTSA Complaints API — for the owner-reported complaint counts shown on model pages, including crash, fire, injury, and death indicators flagged by the submitter.
• NHTSA VPIC (Vehicle Product Information Catalog) API — for VIN decoding. When you enter a VIN, we pass it directly to VPIC to return make, model, year, engine, body class, drivetrain, and other decoded attributes.
• NHTSA VIN Recalls API— for the live "does this specific VIN have an open recall" answer on our VIN results page.

None of these APIs require payment. They are the same underlying data sources that dealerships, insurers, and government-facing lookup tools rely on.

2. How we refresh the data

Our ingestion pipeline runs on a daily cron schedule and pulls the full recall and complaint record set for every brand and model we track. The pipeline writes to a structured database, which the live site reads on demand. Model pages and brand pages are served from that database and cached for up to one hour before re-validating.

VIN lookups, by contrast, are notcached. When you type a VIN into the checker, the request goes straight to NHTSA's live API so that the answer always reflects the most recent data the federal government has about that specific vehicle.

There is an unavoidable small lag — usually under 24 hours — between when NHTSA publishes a new recall and when it appears on our brand and model listing pages. If you need the absolute current state of a specific VIN, the VIN checker is the right tool for that.

3. What we add on top of the raw data

We don't rewrite or edit the underlying NHTSA campaign records. The campaign text, consequence language, and remedy description that appear in each recall card are exactly as NHTSA and the manufacturer published them, minus obvious formatting cruft.

What we do add is a layer of organization and analysis:

• Brand and model grouping— NHTSA's own site makes it surprisingly hard to see every recall a single model has had over time. We pre-compute the full list per brand and per model so you can see it at a glance.
• Component categorization— NHTSA returns a free-text "Component" field that's inconsistent across campaigns. We map it into a small set of broader buckets (brakes, airbags, electrical, engine, software, etc.) so that patterns show up.
• Critical-recall flagging— we scan each campaign's consequence and summary text for keywords related to fire, crash, injury, death, or loss of control and surface the matching campaigns first. It's a simple text filter, not a risk score, and we label it as such.
• Reliability scorecards — on model pages we display a 1–10 score derived from recall count, complaint volume, and the presence of crash, fire, and fatality reports. The formula is documented in our codebase and is meant as a rough comparative signal, not a replacement for a pre-purchase inspection.
• Editorial analysis— some pages include a plain-language analysis section written from the live dataset ("the most common recall category for this brand is X, appearing in Y campaigns"). Those sections are clearly separated from the raw campaign records and exist to put the numbers in context.
• Monthly recap reports— on the blog we publish a monthly summary of every recall that landed in NHTSA's database that month, with a short editorial intro explaining what stood out.

4. What we deliberately don't do

• We don't invent recall campaigns, fabricate campaign numbers, or fill in missing NHTSA data with guesses.
• We don't accept payment from manufacturers to suppress, de-rank, or hide specific recalls.
• We don't store the VINs you look up. VIN lookups are passed through to NHTSA and not retained.
• We don't present owner complaints as defects. Complaints are self-reported and clearly labeled as such.
• We don't give individualized legal, mechanical, or insurance advice. See our disclaimer.

5. Corrections policy

If you believe a page on RecallScanner is wrong — a missing recall, a stale status, a misclassified component, or an incorrect editorial claim — please contact uswith the page URL and a description of the problem. For campaign-level data, we'll verify against the live NHTSA API before making changes; if the issue is in our editorial layer, we'll fix it directly and note the correction in the page's revision history where appropriate.

6. Who runs this site

RecallScanner is built and maintained by a small, independent editorial team with a background in data engineering and automotive research. We are not funded by any automaker, law firm, insurance company, or dealership network. The site covers its operating costs through contextual advertising that is clearly labeled and does not influence which recalls are shown or how they're categorized.