Encar Cars Scraper (South Korea) avatar

Encar Cars Scraper (South Korea)

Pricing

from $1.75 / 1,000 results

Go to Apify Store
Encar Cars Scraper (South Korea)

Encar Cars Scraper (South Korea)

Scrapes used-car listings from Encar, South Korea's largest car marketplace. Each row carries price, mileage, model year, manufacturer, trust/inspection badges and photos, with optional accident history, full options and VIN.

Pricing

from $1.75 / 1,000 results

Rating

0.0

(0)

Developer

Ibnu Adzim

Ibnu Adzim

Maintained by Community

Actor stats

0

Bookmarked

2

Total users

1

Monthly active users

2 days ago

Last modified

Categories

Share

Encar Cars Scraper — South Korea

Scrapes used-car listings from Encar, South Korea's largest used-car marketplace. Domestic and imported vehicles, with optional accident/inspection history, full options and VIN.

  • Domestic (Korean-brand) inventory: ~150,900 listings — Hyundai, Kia, Genesis, Chevrolet/GM Daewoo, KG Mobility/Ssangyong, Renault Korea
  • Imported inventory: ~66,400 listings — Mercedes-Benz, BMW, Audi, Porsche, Jeep, Land Rover and other foreign brands

What each row contains

Every listing is emitted as upstream returns it. Highlights:

  • Price (in 10,000 KRW units), mileage, model year
  • Manufacturer / model / trim (in Korean)
  • Trust and condition badges — extended warranty, home service, inspection report, accident record, resume
  • Photos — all image URLs
  • With includeCarDetails: accident / inspection / seizure history, full options list, VIN, seller contact and dealer information

Record types

recordTypeOne perPurpose
SEARCH_SUMMARYqueryupstream's own count, requests spent, honesty flags
CARlistingthe listing itself, upstream shape preserved
ERRORfailed inputso every input maps to at least one row

An unusually honest API

Most targets in this portfolio hide a bad filter behind a healthy-looking 200 response with the wrong data. Encar's search API does not — an unrecognised manufacturer, or a price/year range with no matches, returns a clean Count: 0 rather than the unfiltered baseline. This Actor still validates carType and sort up front (the API rejects those with an honest HTTP 400, so there's no reason to send a value known to fail), but a mistyped manufacturer costs you nothing but an empty result.

The one real trap: a hidden pagination ceiling

Past roughly offset 10,000, the API does not error and does not clamp back to page 1 — it returns HTTP 200 with listings drawn from a small, non-monotonic pool, which would silently corrupt a naive crawl of a large category. This Actor enforces a hard cap well clear of that boundary and reports resultWindowCeilingHit: true on the summary row when a query has more matches than it can reach, rather than crawling into the degenerate zone. Narrow by manufacturer, price or year to reach the rest of a large category.


Input

FieldNotes
carTypedomestic or import
manufacturerKorean-script brand name, e.g. 현대, 기아, BMW. Run once without it to read exact spellings off results
minPrice / maxPriceIn units of 10,000 KRW
minYear / maxYearYYYYMM format
sortRecently updated, price, or mileage, ascending or descending
queriesOptional array to run several searches in one call
includeCarDetailsOne extra request per listing
maxItemsPer query

Example

{
"carType": "domestic",
"manufacturer": "현대",
"minPrice": 2000,
"maxPrice": 5000,
"minYear": 202001,
"sort": "price-asc",
"includeCarDetails": true,
"maxItems": 100
}

Known limits

  • Pagination cannot reliably reach past ~10,000 results in a single query (see above). Split large categories by manufacturer, price band or year.
  • Price/Year (without Asc/Desc) are not valid sort values — the API rejects them, so only the five options in the input schema are offered.
  • Manufacturer names must be in Korean script as Encar itself uses them; no English-to-Korean translation is performed.

Anti-bot

None. Seven TLS fingerprints tested on both the search and detail endpoints returned clean data every time, and the API even answers requests with no Accept-Language header. The Actor still rotates fingerprints on transient failure and defaults to residential proxy, since cloud egress can be fingerprinted differently from a local test.