Datasetpublic

joakes90/vpic-database

NHTSA vPIC Curated Vehicle Database Curated vehicle database derived from NHTSA's vPIC data, optimized for VIN decoding via the vin-decode-mcp server. Source Data sourced from the National Highway Traffic Safety Administration (NHTSA), an agency of the United States Department of Transportation. NHTSA is a government agency and the services provided are free for use by the public as part of their Open Data initiative. No API key or registration required. Download… See the full description on the dataset page: https://huggingface.co/datasets/joakes90/vpic-database.

sourceHugging Faceotherupdated 1mo agoView on Hugging Face

0likes107downloads

Dataset Card

NHTSA vPIC Curated Vehicle Database

Curated vehicle database derived from NHTSA's vPIC data, optimized for VIN decoding via the vin-decode-mcp server.

Source

Data sourced from the National Highway Traffic Safety Administration (NHTSA), an agency of the United States Department of Transportation. NHTSA is a government agency and the services provided are free for use by the public as part of their Open Data initiative. No API key or registration required.

[Download NHTSA vPIC standalone databases](https://vpic.nhtsa.dot.gov/Downloads/)

What's Included

This is the full vPIC SQLite database with all NHTSA vehicle specification tables, including:

Category	Description
`make`	Vehicle manufacturers (~12,328)
`model`	Vehicle models (~31,869)
`vehicletype`	Vehicle type categories
`make_model`	Make/model relationships
`wmi`	World Manufacturer Identifiers (~12,971)
`wmi_vinschema`	WMI to VIN schema mapping
`pattern`	VIN descriptor patterns (~1.67M)
`vinexception`	VIN exception rules (~18,805)
`vinschema`	VIN schema definitions
`model_years`	Production year ranges
`engine_*`	Engine specifications (model, model, pattern, etc.)
`body_*`	Body specifications (style, cab, type, etc.)
`fuel_*`	Fuel specifications (type, delivery, tank, etc.)
`safety_*`	Safety equipment (airbags, brakes, ABS, etc.)
`battery_*`	Battery/electrification specs
`transmission`	Transmission types
`drivetype`	Drive types (FWD, RWD, AWD, etc.)

Database Stats

—Size: ~627 MB
—Format: SQLite (.db)
—Source vintage: 2026-08

Dataset Viewer

This dataset is available in the Dataset Viewer with individual tables available as separate configurations (config dropdown at top right). Key tables are also exported as Parquet for fast browser-based browsing.

Available Tables (Parquet exports)

Table	Config	Rows	Parquet Size
make	`make`	12,328	0.3 MB
model	`model`	31,869	0.6 MB
make_model	`make_model`	31,869	0.6 MB
wmi	`wmi`	12,971	0.6 MB
pattern	`pattern`	1,674,161	28.6 MB
vehicletype	`vehicletype`	9	<1 MB
wmi_vinschema	`wmi_vinschema`	41,708	0.7 MB
vinexception	`vinexception`	18,805	0.2 MB
vinschema	`vinschema`	25,149	1.4 MB
manufacturer	`manufacturer`	22,893	0.5 MB
country	`country`	199	<1 MB
drivetype	`drivetype`	23	<1 MB
fueltype	`fueltype`	14	<1 MB
bodystyle	`bodystyle`	71	<1 MB
transmission	`transmission`	12	<1 MB
enginemodel	`enginemodel`	346	<1 MB
enginemodelpattern	`enginemodelpattern`	2,510	<1 MB
vspecschemapattern	`vspecschemapattern`	16,754	0.1 MB

Usage

Download the .db file and use with the MCP server:

bash

export VIN_MCP_DB_PATH=/path/to/vPIC_sqlite.db
vin-decode-mcp

Or query directly:

python

import sqlite3
conn = sqlite3.connect("vPIC_sqlite.db", uri=True)
conn.row_factory = sqlite3.Row

# List makes
makes = conn.execute("SELECT * FROM make ORDER BY name").fetchall()

Refresh Frequency

This database is rebuilt from NHTSA's standalone databases approximately every 6-12 months. Check the build timestamp in dataset_info for the source vintage.

Attribution

Vehicle data: NHTSA vPIC Curated database pipeline: vin-decode-mcp