Our data sources, scoring algorithm, and AI inspection process — explained transparently.
AutoSavvy's Deal Score (0–100) estimates how well a used car listing is priced relative to current market conditions. Higher is better: 70+ is a good deal, 40–69 is fair, below 40 means you're likely overpaying.
We compare the asking price to 5–15 similar listings with matching year, make, model, trim, and mileage range in the same geographic market. The gap between asking price and median comparable price is the largest factor.
Mileage is adjusted relative to vehicle age. A 2020 car with 90,000 miles is a different value proposition than one with 30,000. We use age-normalized mileage curves to assess wear.
Known issues, recall history, complaints data, and listing details (number of owners, accident history mentions) adjust the score. Photo damage detection (if photos provided) can reduce scores by up to 30 points.
Pricing varies significantly by geography. A truck in rural Texas and the same truck in Manhattan have different market values. We apply regional demand adjustments based on listing location.
We decode VINs using the NHTSA Vehicle Product Information Catalog (vPIC) API — a free U.S. government service. This gives us authoritative vehicle specifications: exact trim, engine, drivetrain, and equipment packages. Free and no rate limits.
Owner-reported complaints submitted to the National Highway Traffic Safety Administration. We surface the top complaint categories for a vehicle's year/make/model to flag recurring reliability issues. Data is publicly available at nhtsa.gov.
We check for open safety recalls using the NHTSA Recalls API. Recalls that haven't been remediated are a material fact about a vehicle's condition and are included in the full Vehicle Intelligence Report.
We source real-time market listings of similar vehicles to build pricing comparables. These are actual listings at dealers and private sellers in the relevant geographic market, filtered by year, make, model, trim, and mileage band. Comparables update as market conditions change.
When you upload vehicle photos, we send them to AI Vision (OpenAI) for damage analysis. The AI looks for:
Detected damage adjusts the Deal Score downward by up to 30 points, depending on severity and number of issues found. The photo inspection is an AI estimate — it is not a substitute for an in-person mechanical inspection by a qualified technician.
We use AI to analyze listings. Scores update as market data changes. No dealer relationships influence our scores. We are not affiliated with any dealer, manufacturer, or car listing platform. Our only incentive is to give you accurate pricing intelligence.
Enter a VIN, paste a listing URL, or upload a screenshot. Get your Deal Score in 30 seconds.
Score a Deal Free →