Naver Cafe Search Results Scraper

Search public Naver Cafe results by keyword and collect structured article-card data. The Actor uses Naver's Cafe search tab and keeps the search-page evidence together with each result, including article and Cafe URLs, the displayed date, excerpt, optional thumbnail, and reply snippets visible on the card. It only reads public search cards; it does not open article pages, sign in, or collect private or members-only material.

What it returns

Each result has 31 possible schema fields. Some fields are optional because not every Cafe card displays them.

Field group	Fields
Result identity	recordType, id, articleKey, articleId, cafeSlug, position, queryPosition, page, title, url, link, canonicalLink
Cafe	cafeId, cafeName, cafeUrl, source, sourceDomain
Visible card content	snippet, description, dateRaw, publishedAtText, publishedAt, thumbnailUrl, replyPreviews, replyPreviewCount
Search provenance	query, searchUrl, pageNumber, pageStart, sortOrder, scrapedAt

publishedAtText retains the page's original date label; publishedAt is present when that label is an absolute date. replyPreviewCount counts only the visible snippets captured on the card. It is not a count of every reply to the article. replyPreviews is capped at five entries. thumbnailUrl is omitted when the page has no result thumbnail or no usable URL.

Input

query and queries can be used together; blank and duplicate batch entries are removed. If both are omitted, the runtime query is 인공지능. maxItems is a global cap from 1 to 200. maxPages defaults to 5 and accepts 1–20 pages per query; maxConcurrency defaults to 2 and accepts 1–5 browser pages; maxRequestRetries defaults to 2 and accepts 0–5 retries. The optional proxyConfiguration is passed to Apify Proxy when supplied; it does not bypass Naver's access controls. The Actor uses the page's default relevance order, deduplicates by canonical article URL, and advances search offsets by the number of results on the current page.

{
  "query": "technology",
  "queries": ["laptop", "smartphone"],
  "maxItems": 50
}

Small preview

The Console's prefilled input is a small preview: query: 인공지능, maxItems: 5, maxPages: 1, maxConcurrency: 1, maxRequestRetries: 0, and proxyConfiguration: { "useApifyProxy": false }. These are UI prefills only. API calls that omit these fields retain the runtime defaults: 50 items, 5 pages per query, concurrency 2, retries 2, and no proxy configuration. Batch queries and the existing output aliases remain supported.

An empty result page stops pagination without changing the collected count. Extraction failures are reported as errors rather than empty results. Login gates, CAPTCHA/human-verification pages, and HTTP 401/403/407/429 responses fail without retrying that request or solving the gate. The Actor uses normal TLS validation and browser security settings, with no webdriver masking or fingerprint injection.

There are no fixture-file, fixture-data, or debug input controls. Unrecognized input properties are rejected.

Extraction verification

On 2026-10-02, the exact public preview query 인공지능 rendered 30 list-view Cafe cards in the cloud browser and no ugcItem cards. The extractor supports both observed layouts. It reads the article title, snippet, Cafe name, date and visible reply previews from each card, omits missing-image placeholders, and fails if a page has neither usable cards nor an explicit no-results message. Access challenges remain terminal.

npm test includes offline regressions against sanitized excerpts of that public DOM in test/fixtures/cafe-list-view-2026-10-02.html. These tests do not contact Naver or prove that an Actor cloud run succeeded. Cloud runtime verification is separate.

Chrome observation example

The following populated record is an observation of the live Chrome Cafe search page for technology, not a result from an Actor run. Chrome first loaded the Actor's search route, then the visible Cafe tab was selected; Naver showed the Cafe URL below, the relevance sort, 60 result cards, and the article at position 21. The displayed card supplied its Cafe name, date, title, excerpt, URL, and two reply previews. The thumbnail itself was visible in Chrome, but its URL was not exposed in the accessibility observation, so no thumbnail URL is guessed here.

{
  "recordType": "cafe_search_result",
  "position": 21,
  "id": "songdolovemom:7048",
  "articleKey": "7048",
  "cafeSlug": "songdolovemom",
  "queryPosition": 21,
  "page": 1,
  "title": "마이크론 테크놀로지 주가 (Micron Technology, MU) 급등 이유",
  "url": "https://cafe.naver.com/songdolovemom/7048",
  "link": "https://cafe.naver.com/songdolovemom/7048",
  "canonicalLink": "https://cafe.naver.com/songdolovemom/7048",
  "snippet": "마이크론 테크놀로지 주가 급등 이유 (Micron Technology, MU) 마이크론 테크놀로지는 2025년 말부터 2026년 들어 AI 메모리 붐의 최대 수혜주로 부상하며 주가가 폭발적으로 상승했습니다. 2026년 들어 YTD(연초 대비) 70~150% 이상 상승하는 등 반도체 섹터에서 가장 강력한 랠리를 보여주고 있으며, 일부 기간에는 300%를 넘는 누적 상승률을 기록하기도 했습니다. AI 수요 폭발과 HBM(고대역폭 메모리)의...",
  "description": "마이크론 테크놀로지 주가 급등 이유 (Micron Technology, MU) 마이크론 테크놀로지는 2025년 말부터 2026년 들어 AI 메모리 붐의 최대 수혜주로 부상하며 주가가 폭발적으로 상승했습니다. 2026년 들어 YTD(연초 대비) 70~150% 이상 상승하는 등 반도체 섹터에서 가장 강력한 랠리를 보여주고 있으며, 일부 기간에는 300%를 넘는 누적 상승률을 기록하기도 했습니다. AI 수요 폭발과 HBM(고대역폭 메모리)의...",
  "cafeId": "songdolovemom",
  "cafeName": "인천&송도 생활정보",
  "cafeUrl": "https://cafe.naver.com/songdolovemom",
  "articleId": "7048",
  "dateRaw": "2026.05.26.",
  "publishedAtText": "2026.05.26.",
  "publishedAt": "2026-05-26",
  "replyPreviews": [
    "반도체는 주가가 여전히 좋은가봐요. 좋은 정보 감사합니다~",
    "오늘 미국주식 반도체와 우주섹터는 다 오르는군요."
  ],
  "replyPreviewCount": 2,
  "query": "technology",
  "searchUrl": "https://search.naver.com/search.naver?ssc=tab.cafe.all&sm=tab_jum&query=technology",
  "pageNumber": 1,
  "pageStart": 1,
  "sortOrder": "relevance",
  "source": "Naver Cafe",
  "sourceDomain": "cafe.naver.com",
  "scrapedAt": "2026-09-22T20:19:25.000Z"
}

Downloading and use

You can download the dataset in various formats such as JSON, HTML, CSV, or Excel. The result page may change by query, locale, availability, and time; the Actor preserves the fields Naver exposes rather than filling absent data with guesses.

Each page requires a browser navigation. Larger maxItems values can require more page loads and increase runtime and compute usage; optional proxy routing can also affect run cost. There is no fixed price claim here: review the run details in Apify Console for the current compute and proxy usage.

If a run returns fewer results than requested, check that the public Cafe results exist for the query and inspect the run log for an empty page or repeated results. For a schema or extraction issue, use the Actor's Issues tab and include the run ID and a short description. This Actor is not affiliated with Naver. Use it only for public information, follow Naver's terms and applicable law, respect rate limits, and do not attempt to access private or members-only Cafe content. Public posts may still contain personal data; collect and use it only when you have a lawful basis.