What Is a VIN (Vehicle Identification Number)?

Every vehicle manufactured for sale in the United States since 1981 has a 17-character Vehicle Identification Number. It's stamped into the dashboard (visible through the windshield on the driver's side), printed on the driver's door jamb sticker, and recorded on the title, registration, and insurance documents.

The VIN isn't random. Each character encodes specific information about the vehicle's origin, manufacturer, model, engine, and production sequence.

How to read a VIN

Characters 1–3 are the World Manufacturer Identifier (WMI). Character 1 is the country of manufacture (1 = US, J = Japan, W = Germany). Characters 2–3 identify the manufacturer and vehicle type.

Characters 4–8 are the Vehicle Descriptor Section (VDS). These encode the model, body type, engine, and transmission. This is where most of the spec information lives, but the encoding is manufacturer-specific — position 5 might mean "engine type" for Toyota and "body style" for Ford.

Character 9 is a check digit. It's a mathematical validation that the rest of the VIN is self-consistent. If someone transcribes a VIN wrong, the check digit will usually catch it.

Character 10 is the model year. A = 2010, B = 2011, continuing through the alphabet (skipping I, O, Q, U, Z) and then cycling through numerics.

Character 11 is the assembly plant.

Characters 12–17 are the sequential production number.

Why VIN decoding is harder than it looks

NHTSA maintains the vPIC database, which is the closest thing to a universal VIN decoder. Manufacturers submit their VIN encoding patterns to NHTSA via 565 submissions. But these submissions are inconsistent. Some manufacturers provide full trim-level resolution. Others only decode to the model level, leaving out critical details like engine size or drivetrain configuration.

This means that decoding a VIN through vPIC might tell you "2020 Honda Civic" without telling you whether it's the 158 hp LX or the 306 hp Type R. For some manufacturers, you get full detail. For others, you get a partial answer. You won't know which until you try.

Commercial VIN decoding services fill these gaps by maintaining their own databases of VIN patterns, but they charge per-decode and their accuracy varies by manufacturer and model year.

When you have a VIN vs. when you don't

VIN decoding is the right approach when you're working with a specific physical vehicle — an insurance claim, a repair order, a title transfer. You have the VIN, you need the specs.

Year/make/model lookup is the right approach when you're building a catalog, a comparison tool, or a search interface. You don't have a VIN — you have a user who selected "2024 Ford Mustang GT" from a dropdown.

Most applications need both paths. CarVector focuses on the YMM path — structured spec data by year, make, model, and trim. For VIN decoding, NHTSA's free vPIC API is the starting point, though the trim resolution gaps mean you may need to supplement with a commercial decoder.