Vehicle Reference Scraper

This is a vehicle SPECIFICATION / SAFETY / RECALL reference source. It is NOT a for-sale vehicle listings source. There are no prices, no mileage, no dealers and no inventory here. Every record describes a model or a VIN — its factory specs, its EPA fuel economy, its NHTSA crash-test stars and its recall history. If you need cars currently on sale, use a marketplace actor instead.

Data comes from three public US government web services, all keyless:

Source	What it gives
NHTSA vPIC — vpic.nhtsa.dot.gov/api/vehicles	VIN decoding (make, model, trim, engine, body, plant, safety equipment), make and model lists per model year
NHTSA — api.nhtsa.gov	5-Star Safety Ratings (overall / front / side / rollover), recall campaigns, owner complaint counts, investigation counts
fueleconomy.gov — www.fueleconomy.gov/ws/rest/vehicle	EPA city / highway / combined MPG, annual fuel cost, CO2 tailpipe grams per mile, greenhouse-gas and smog scores, drivetrain, transmission, EPA vehicle class

No API key, no account and no sign-up. The actor runs against the live services on every run.

Modes

mode	Needs	Returns
model	year, make, model	One record per EPA fuel-economy configuration (engine / transmission combination) for that vehicle, each enriched with the NHTSA safety rating, recall count and complaint count
vin	vins	One record per VIN — the full vPIC decode, enriched with the safety rating, recall count and complaint count for the year/make/model the VIN resolves to
makes	year (optional)	One record per make. With a year, the makes fueleconomy.gov lists for that model year; without a year, the full NHTSA vPIC make list
models	year, make	One record per model NHTSA vPIC lists for that make and model year, with numeric make and model IDs

Set include_safety or include_recalls to false to skip those extra lookups and run faster.

Output

One flat snake_case record per vehicle (per VIN in vin mode, per fueleconomy.gov vehicle ID in model mode), plus a raw field holding the untouched upstream payload. All records from one mode carry the same keys; the samples below leave out the ones that are empty in that mode.

Sample — mode: "model", year 2020, make Honda, model Accord

{
  "record_type": "model",
  "source": "fueleconomy_gov",
  "source_url": "https://www.fueleconomy.gov/feg/Find.do?action=sbs&id=42071",
  "model_year": 2020,
  "make": "Honda",
  "model": "Accord",
  "epa_configuration": "Auto (AV-S7), 4 cyl, 1.5 L, Turbo",
  "drive_type": "Front-Wheel Drive",
  "engine_cylinders": 4,
  "displacement_l": 1.5,
  "engine_descriptor": "SIDI",
  "fuel_type_primary": "Regular Gasoline",
  "transmission_style": "Automatic (AV-S7)",
  "fueleconomy_id": 42071,
  "epa_vehicle_class": "Large Cars",
  "mpg_city": 29.0,
  "mpg_highway": 35.0,
  "mpg_combined": 31.0,
  "annual_fuel_cost_usd": 2100.0,
  "co2_tailpipe_gpm": 282.0,
  "ghg_score": 7,
  "smog_score": 7,
  "is_turbocharged": true,
  "is_supercharged": false,
  "start_stop": "N",
  "nhtsa_model": "ACCORD",
  "safety_vehicle_id": 14818,
  "safety_vehicle_description": "2020 Honda ACCORD 4 DR FWD",
  "safety_overall_rating": 5,
  "safety_front_crash_rating": 5,
  "safety_side_crash_rating": 5,
  "safety_rollover_rating": 5,
  "safety_rollover_possibility": 0.093,
  "nhtsa_electronic_stability_control": "Standard",
  "nhtsa_forward_collision_warning": "Standard",
  "nhtsa_lane_departure_warning": "Standard",
  "recall_count": 5,
  "latest_recall_campaign": "26V332000",
  "latest_recall_component": "AIR BAGS:SENSOR:OCCUPANT CLASSIFICATION:FRONT PASSENGER",
  "complaint_count": 362,
  "investigation_count": 1
}

Sample — mode: "vin", VIN 5YJ3E1EA6PF384836

{
  "record_type": "vin",
  "source": "nhtsa_vpic",
  "source_url": "https://vpic.nhtsa.dot.gov/api/vehicles/DecodeVinValues/5YJ3E1EA6PF384836?format=json",
  "vin": "5YJ3E1EA6PF384836",
  "vin_pattern": "5YJ3E1EA*PF",
  "model_year": 2023,
  "make": "TESLA",
  "model": "Model 3",
  "manufacturer": "TESLA, INC.",
  "vehicle_type": "PASSENGER CAR",
  "body_class": "Sedan/Saloon",
  "doors": 4,
  "seats": 5,
  "seat_rows": 2,
  "fuel_type_primary": "Electric",
  "electrification_level": "BEV (Battery Electric Vehicle)",
  "battery_type": "Lithium-Ion/Li-Ion",
  "transmission_style": "Automatic",
  "transmission_speeds": 1,
  "gvwr": "Class 1C: 4,001 - 5,000 lb (1,814 - 2,268 kg)",
  "curb_weight_lb": 4048,
  "plant_city": "FREMONT",
  "plant_state": "CALIFORNIA",
  "plant_country": "UNITED STATES (USA)",
  "abs": "Standard",
  "esc": "Standard",
  "tpms": "Direct",
  "backup_camera": "Standard",
  "blind_spot_monitor": "Standard",
  "air_bag_front": "1st Row (Driver and Passenger)",
  "air_bag_side": "1st and 2nd Rows",
  "decode_status_code": "0",
  "decode_status_text": "0 - VIN decoded clean. Check Digit (9th position) is correct",
  "nhtsa_model": "MODEL 3",
  "recall_count": 12,
  "latest_recall_campaign": "26V507000",
  "latest_recall_component": "EXTERIOR LIGHTING:HEADLIGHTS",
  "complaint_count": 425
}

Sample — mode: "models", year 2020, make Honda

{
  "record_type": "model_name",
  "source": "nhtsa_vpic",
  "source_url": "https://vpic.nhtsa.dot.gov/api/vehicles/GetModelsForMakeYear/make/Honda/modelyear/2020?format=json",
  "model_year": 2020,
  "make": "HONDA",
  "make_id": 474,
  "model": "Accord",
  "model_id": 1861
}

Notes on the data

• safety_* fields are populated only where NHTSA actually crash-tested that year/make/model. Older or unrated vehicles come back with null stars even though the rest of the record is complete — that is NHTSA's own coverage, not a scrape failure.
• mode: "model" returns one row per drivetrain configuration, so a single nameplate typically yields several rows (the 2020 Accord returns five).
• make_id is null in mode: "makes" when the year-scoped make list comes from fueleconomy.gov, which does not publish numeric make IDs. Run makes without a year to get NHTSA vPIC make IDs instead.
• Model names differ between the two agencies. NHTSA uses Accord; fueleconomy.gov also lists Accord Hybrid and Accord 2.0T Sport/Touring as separate models, and folds drive and trim into names such as RAV4 AWD LE. Use mode: "models" to see NHTSA's names. nhtsa_model is the NHTSA model the safety, recall and complaint fields were looked up under (RAV4 for RAV4 AWD LE). When NHTSA has no model of that name for the make and year, nhtsa_model is empty and recall_count / complaint_count are null (unknown), not 0.
• NHTSA rates each body and drivetrain separately (for example

  2020 Toyota CAMRY 4 DR FWD
  and ... 4 DR AWD). The safety_* fields come from the variant whose drive (and door count, where known) matches the record, and safety_vehicle_description names it. When the record does not say enough to pick one variant — common in VIN decodes, which often omit the drive type — the safety_* fields are left empty rather than filled from a guessed variant.
• latest_recall_campaign / latest_recall_component are the campaign NHTSA received most recently for that year/make/model.
• trim is the VIN-decoded trim and is only filled in vin mode; fueleconomy.gov publishes no trim. In model mode epa_configuration holds the EPA's label for the engine / transmission configuration (for example Auto (AV-S7), 4 cyl, 1.5 L, Turbo).
• engine_model is the manufacturer engine code from the VIN decode (vin mode). engine_descriptor is the EPA engine technology note from fueleconomy.gov (model mode), e.g. SIDI (direct injection), PHEV, FFV.
• electrification_level is filled only for electrified vehicles (hybrid, plug-in hybrid, battery-electric, fuel-cell); diesel, flex-fuel and CNG vehicles leave it empty and show their fuel in fuel_type_primary.
• mpg_city / mpg_highway / mpg_combined are the EPA figures on the primary fuel: miles per gallon-equivalent (MPGe) for battery-electric vehicles and the gasoline-only figure for plug-in hybrids. electric_range_miles is the EPA all-electric range (the electric-only range for plug-in hybrids).
• ghg_score and smog_score are EPA scores from 1 (worst) to 10 (best). safety_rollover_possibility is NHTSA's rollover risk as a fraction (0.093 = 9.3%).

Proxy

proxyConfiguration is optional and off by default — these are public government services. If your network cannot reach them, set a US proxy and it is applied to every request the actor makes. A run that cannot reach a service fails with the reason rather than returning partial or placeholder rows.

Input example

{
  "mode": "model",
  "year": "2020",
  "make": "Honda",
  "model": "Accord",
  "include_safety": true,
  "include_recalls": true,
  "limit": 20
}