Daangn Scraper - Karrot Used Price Index (Korea Second-hand)

What does it sell for used in Korea? This Daangn scraper (also searched as a Karrot scraper) turns Daangn Market (당근마켓, "Karrot") public search results into a price index of Korean used prices: median, P10 to P90 percentiles and an outlier-trimmed mean of second-hand asking prices per keyword and per neighborhood. Relevance filters remove accessories, parts, wanted ads and placeholder prices before the stats are computed.

Live-tested on Apify: YES (2026-10-29). "아이폰15" gave a median of 630,000 KRW and "닌텐도 스위치" 200,000 KRW from the public search page.

Why this actor

• An answer, not a pile of listings: one row per keyword and region, with clean statistics.
• Relevance filters that matter on Daangn: the search is fuzzy (searching 아이폰15 also returns iPhone 6S and cases). Title-must-match, exclude words, a minimum price and a below-median filter keep the stats honest.
• Privacy by design: no seller names or profiles, no listing descriptions (they often contain phone numbers and addresses), no coordinates and no photos. Optional listing rows contain only title, price, status, neighborhood, posted time and URL.
• Polite: respects robots.txt, 1 request every ~4 s, a single concurrent request.

Input

{ "keywords": ["아이폰15", "닌텐도 스위치"], "regions": ["반포본동-368"], "excludeWords": ["케이스", "부품용", "삽니다"], "outputListings": false }

regions are Daangn region slugs from the website URL (?in=반포본동-368). Leave it empty for Daangn's default results.

Output (price_index row)

Example (illustrative values). Out of 5 listings found, the filters kept 2: a case, a "부품용" (for parts) phone and a free giveaway were dropped.

{
  "recordType": "price_index", "keyword": "아이폰15", "region": null,
  "listingsFound": 5, "listingsUsedForStats": 2,
  "all": { "count": 2, "min": 570000, "p10": 574000, "p25": 580000, "median": 590000, "p75": 600000, "p90": 606000, "max": 610000, "mean": 590000, "trimmedMean": 590000, "outliersRemoved": 0 },
  "onSale": { "count": 1, "min": 570000, "p10": 570000, "p25": 570000, "median": 570000, "p75": 570000, "p90": 570000, "max": 570000, "mean": 570000, "trimmedMean": 570000, "outliersRemoved": 0 },
  "sold": { "count": 1, "min": 610000, "p10": 610000, "p25": 610000, "median": 610000, "p75": 610000, "p90": 610000, "max": 610000, "mean": 610000, "trimmedMean": 610000, "outliersRemoved": 0 },
  "currency": "KRW",
  "sourceUrl": "https://www.daangn.com/kr/search/buy-sell/?q=%EC%95%84%EC%9D%B4%ED%8F%B015",
  "fetchedAt": "2026-10-29T06:00:00.000Z"
}

onSale covers listings still for sale; sold covers sold and reserved ones (on live pages it is usually empty, see Limitations). With outputListings: true you also get one listing row per listing (title, priceKrw, status, neighborhood, postedAt, url).

Pages used

• https://www.daangn.com/kr/search/buy-sell/?q={keyword}&in={region}: public, no login, allowed by robots.txt for generic user agents. Data is read from the page's embedded JSON (__remixContext), with the schema.org ld+json ItemList as a fallback.

Limitations

• These are asking prices, not transaction prices. Daangn's public search shows on-sale listings only, so sold is usually empty.
• One search page shows about 270 to 300 listings. There is no deeper pagination on the public page.
• The keyword's "region" depends on Daangn's own location logic. Results may include nearby neighborhoods.
• Daangn could change its page structure. Parsing has a fallback, and the run warns if 0 listings are parsed.

API keys

None. This actor reads public pages only, takes no credentials and stores none.

Pricing (pay per event)

• price-index: $30.00 per 1,000 index rows ($0.03 per row). One row is one keyword in one region and summarizes up to ~300 listings.
• listing: $1.00 per 1,000 listings (only if outputListings is on).
• Actor start: Apify's synthetic apify-actor-start event, a fraction of a cent per run.
• Platform usage is included. You pay only the event prices above; the Apify compute of the run is not billed to you separately.
• Set Max total charge for a run to cap spend; the actor stops early when it is reached.

FAQ

Do I need an API key or a Daangn account? No. The actor reads Daangn's public search page, with no login and no key.

How fresh is the data? Each run loads the search page live, so the index reflects the listings shown on Daangn at that moment. Nothing is cached between runs.

Can I schedule it daily? Yes. An Apify Schedule with the same keywords gives you one index row per keyword and region per day, which you can chart as a used-price trend. Note that each run is a fresh snapshot; the actor does not compare with the previous run.

What does a typical run cost? One keyword costs $0.03. 10 keywords in 3 regions (30 index rows) cost about $0.90. Adding raw listings for one keyword (about 300 rows) adds about $0.30.

Is this legal, and where does the data come from? The data comes from Daangn's public search pages, which robots.txt allows for generic crawlers. The actor does not log in, collects no seller data, and crawls slowly. It is not affiliated with Daangn (Karrot); you are responsible for how you use the results.