# Naver News Scraper (네이버 뉴스 · Search, Full Text, Rankings) (`magenta_courser/naver-news-scraper`) Actor

Enter keywords or article URLs and get Naver News (네이버 뉴스) articles as flat rows: title, publisher, publish and update time, Naver News URL, original URL, thumbnail, section and optional full text. Also section feeds, publisher lists, rankings and a new-articles-only monitoring mode.

- **URL**: https://apify.com/magenta_courser/naver-news-scraper.md
- **Developed by:** [SUNGHWAN CHO](https://apify.com/magenta_courser) (community)
- **Categories:** News, Marketing
- **Features:** Limited permissions
- **Stats:** 2 total users, 1 monthly users, 100.0% runs succeeded, 0 bookmarks
- **User rating**: No ratings yet

## Pricing

from $0.32 / 1,000 articles

This Actor is paid per event. You are not charged for the Apify platform usage, but only a fixed price for specific events.
Since this Actor supports Apify Store discounts, the price gets lower the higher subscription plan you have.

Learn more: https://docs.apify.com/actors/running/actors-in-store.md#pay-per-event

## What's an Apify Actor?

An Actor is a serverless cloud program that runs on the Apify platform. It has two run modes.
In Batch mode, an Actor accepts a well-defined JSON input, performs an action which can take anything from a few seconds to a few hours,
and optionally produces a well-defined JSON output, datasets with results, or files in key-value store.
In Standby mode, an Actor provides a web server which can be used as a website, API, or an MCP server.

Apify vocabulary and the platform model are defined once, in the agent quickstart at https://apify.com/agents.md.

## How to integrate an Actor?

If asked about integration, you help developers integrate Actors into their projects.
You adapt to their stack and deliver integrations that are safe, well-documented, and production-ready.

Do not guess an integration path. Every one of them is in the agent quickstart at https://apify.com/agents.md: the Apify MCP server, Agent Skills with the Apify CLI, the JavaScript and Python clients, the REST API, and the account-free path for an agent with no human to sign in. It also carries the rule on stating cost before the first paid run.

For examples already wired to this Actor's own input schema, see the [API](#api) section below.

Each client library has reference documentation the quickstart does not restate: [JavaScript/TypeScript](https://docs.apify.com/api/client/js/docs.md) (`npm install apify-client`) and [Python](https://docs.apify.com/api/client/python/docs.md) (`pip install apify-client`).

# README

## Naver News Scraper (네이버 뉴스) | Korean News Search, Full Article Text, Rankings & Monitoring (2026)

**Languages:** English · [한국어](#korean) · [中文](#chinese) · [日本語](#japanese)

[![$0.40 per 1,000 Naver News articles as flat JSON, full article text included](https://raw.githubusercontent.com/experimory-crypto/naver-map-scraper-examples/main/assets/hero-naver-news-scraper.png)](https://apify.com/magenta_courser/naver-news-scraper)

Enter **keywords, article URLs, sections, publisher codes or a ranking** and get **Naver News (네이버 뉴스) articles** as flat JSON rows: title, publisher, publish and update time, Naver News URL, original publisher URL, thumbnail, section and, if you ask for it, the full article text and body images. A monitoring mode returns only the articles you have not received before. No Naver login, no API key, no browser.

Part of a small set of Korean data Actors from one maintainer: [Naver Search](https://apify.com/magenta_courser/naver-search-scraper), [Naver DataLab trends](https://apify.com/magenta_courser/naver-datalab-trends-scraper), [Naver AI Briefing](https://apify.com/magenta_courser/naver-ai-briefing-scraper) and [Naver Map reviews](https://apify.com/magenta_courser/naver-place-reviews-scraper). **Reporter names, comments and comment writers are never read as data fields.**

> **Free plan:** Apify's $5 monthly credit covers about 12,500 articles with this Actor (at the per-article price; each run adds a $0.00005 start fee).

### Why choose this Actor?

**$0.40 per 1,000 articles, full text included.**<br>
You pay $0.0004 per article row saved. Turning on `includeArticleContent` adds the full text, exact times, sections and body images to the same row at no extra charge.

**Five ways in, one row format out.**<br>
Keyword search, article URLs, section feeds, one publisher's article list, and the most-viewed or most-commented rankings all return the same 26 fields, with `foundBy` telling you where the row came from. Several can be combined in one run.

**Built for monitoring, not only one-time exports.**<br>
`onlyNewSinceLastRun` remembers which articles it already returned for each list and returns only new ones. With newest-first sorting it stops reading once it reaches articles older than the previous run, so a quiet keyword reads few pages and returns few or no rows.

**Honest about what Naver News hosts.**<br>
Naver's news search also lists articles that exist only on the publisher's own site. Those rows are returned with `naverNewsUrl: null` and the link in `originalUrl`; their text cannot be read, and the row says so with `contentStatus: "unsupported"`.

**Measured, not estimated.** 10 articles with full text took 17.5 seconds at 256 MB, 10 of 10 texts read (run `VV3yRjyacYH5eofNg`, build 0.1.8, 2026-11-11, default datacenter proxy, no retries). 100 list rows without text took 56.9 seconds (run `kkIzYQnrVM6s7RUtH`, same build).

### Quick start

The 100 newest articles for one keyword, list data only. This is the default input; at the Free-plan price it costs $0.04.

```json
{
  "keywords": ["삼성전자"]
}
```

[Open the Actor input](https://console.apify.com/actors/cpN1GPyl1f65kfEvT/input), type keywords one per line (`삼성전자` means Samsung Electronics), and click **Start**. Export the dataset as JSON, CSV or Excel.

#### Copy to your AI assistant

Paste this into ChatGPT, Claude, Cursor or another assistant so it can call the Actor for you:

```text
Use the Apify Actor magenta_courser/naver-news-scraper to get Naver News (South Korea) articles.

Input (JSON), fill at least one of: "keywords" (list of search keywords, Korean works best, e.g. ["삼성전자"]), "articleUrls" (Naver News article links or "001/0016436704"), "sections" (any of politics, economy, society, culture, world, science), "pressIds" (publisher codes such as "001" for Yonhap News), "ranking" ("mostViewed" or "mostCommented"). Optional: "maxResults" (1-1000 per keyword / section / publisher / ranking, default 100), "sort" ("recent", "relevance", "oldest"), "includeArticleContent" (boolean, adds full text), "daysBack" (integer, last N days in Korea time) or "dateFrom"/"dateTo" ("YYYY-MM-DD"), "onlyNewSinceLastRun" (boolean, for scheduled monitoring) with "historyStoreName".

Output: one dataset row per article with id, foundBy, keyword, rank, title, url, naverNewsUrl, originalUrl, pressId, articleId, pressName, summary, thumbnailUrl, publishedAt (ISO, Korea time), publishedAtPrecision (minute/hour/day/week/month), updatedAt, section, sections, viewCount, commentCount, content, contentLength, imageUrls, contentStatus (null, ok, empty, unsupported, notFound, failed), scrapedAt. Missing values are null. Reporter names and comments are not returned as fields.

Run it through the Apify API: POST https://api.apify.com/v2/acts/magenta_courser~naver-news-scraper/run-sync-get-dataset-items with my Apify token as a Bearer token and the input as the JSON body. Or connect the MCP server https://mcp.apify.com?tools=magenta_courser/naver-news-scraper.

Price: $0.0004 per article row (Free plan), full text included. Keep maxResults small while testing.
```

#### Use it over MCP

Add this server to Claude Desktop, Cursor or any MCP client, authenticated with your Apify token:

```text
https://mcp.apify.com?tools=magenta_courser/naver-news-scraper
```

The agent only has to fill `keywords` (or one of the other four sources); everything else has a default.

### Key features

- **Keyword search.** Newest first, by relevance or oldest first; a fixed date range or the last N days; up to 1,000 rows per keyword.
- **Full article text on request.** `content` is one paragraph per line with photo captions removed, plus `contentLength`, body `imageUrls`, all `sections`, `updatedAt` and the publisher's `originalUrl`.
- **Article URLs in, articles out.** Accepts `n.news.naver.com`, `m.news.naver.com` and `news.naver.com` article links, Naver Sports and Naver Entertainment article links, old `read.naver?oid=..&aid=..` links, or just `001/0016436704`.
- **Section feeds.** Politics, economy, society, culture, world and science: the section's current headlines followed by its article feed.
- **Publisher lists.** One publisher's articles on Naver News, newest first, day by day.
- **Rankings.** Naver's most-viewed or most-commented ranking across publishers, or one publisher's top 20 with the view or comment count Naver shows.
- **Time precision is stated.** Lists often show only "3시간 전"; `publishedAtPrecision` says whether `publishedAt` is exact to the minute, hour, day, week or month. Reading the article page makes it exact to the minute.
- **Stable IDs for de-duplication.** `id` is `naver-news:<pressId>/<articleId>`, or `web:<host><path>` for an article that is only on the publisher site.
- **Incremental monitoring.** `onlyNewSinceLastRun` with a separate history per keyword, section, publisher and ranking.
- **E-mail addresses are redacted.** Addresses in titles, snippets and article text are replaced with `[email redacted]`.
- **A free run summary.** `RUN_SUMMARY` has the status and stop reason of every list, article-text counts, and for `articleUrls` every address that was not returned with the reason.
- **Clean stop at your cost limit.** Rows stored equal rows charged; lists that were cut off or not started are marked `budgetLimited`.

### How to scrape Naver News

#### Latest articles for keywords, last 7 days

`daysBack` is counted in Korea time and moves with the run date, so it stays correct in a schedule.

```json
{
  "keywords": ["삼성전자", "SK하이닉스"],
  "daysBack": 7,
  "maxResults": 300
}
```

#### Articles with full text

One extra page load per article. Rows without a copy on Naver News keep `content: null` and get `contentStatus: "unsupported"`.

```json
{
  "keywords": ["반도체 수출"],
  "sort": "relevance",
  "maxResults": 50,
  "includeArticleContent": true
}
```

(`반도체 수출` = chip exports.)

#### Read specific articles in full

No keyword needed. Each readable article becomes one row with full text. Addresses that cannot be read are listed with the reason in `RUN_SUMMARY` and are not charged.

```json
{
  "articleUrls": [
    "https://n.news.naver.com/mnews/article/015/0005352683",
    "001/0016442797"
  ]
}
```

#### A fixed date range

`dateFrom` and `dateTo` are Korea-time dates and apply to keyword searches and publisher lists. Naver lists about 1,000 results per search, so for a long period run once per shorter range.

```json
{
  "keywords": ["환율"],
  "dateFrom": "2026-10-01",
  "dateTo": "2026-10-31",
  "sort": "oldest",
  "maxResults": 1000
}
```

#### Hourly or daily monitoring: only articles you have not received before

Schedule this input. The first run returns the latest articles as usual; later runs return only new ones.

```json
{
  "keywords": ["올리브영"],
  "onlyNewSinceLastRun": true,
  "historyStoreName": "oliveyoung-news-hourly"
}
```

#### Section feeds

About 10 current headlines of the section (`foundBy: "sectionHeadline"`) followed by its article feed (`foundBy: "section"`).

```json
{
  "sections": ["economy", "science"],
  "maxResults": 50
}
```

#### One publisher's articles

`001` is Yonhap News. Without dates the list covers the last 7 days.

```json
{
  "pressIds": ["001"],
  "daysBack": 1,
  "maxResults": 200
}
```

#### A publisher's most-viewed articles with view counts

Up to 20 rows per publisher, with `viewCount` (or `commentCount` for `mostCommented`).

```json
{
  "pressIds": ["001"],
  "ranking": "mostViewed",
  "maxResults": 20
}
```

#### Naver's ranking across all publishers

The top 5 articles of about 80 publishers, about 400 rows, so raise `maxResults` to receive all of them. `rank` is the rank within the publisher; this page shows no counts.

```json
{
  "ranking": "mostViewed",
  "maxResults": 500
}
```

### Input parameters

Fill at least one of `keywords`, `articleUrls`, `sections`, `pressIds` or `ranking`. Only `keywords` is prefilled in the console.

| Field | Type | Default | Description |
|---|---|---|---|
| `keywords` | array of strings | — | Search keywords, one per line. Each is searched separately. Korean keywords return the most results. |
| `maxResults` | integer | `100` | 1 to 1,000. Maximum rows for each keyword, each section, each publisher and each ranking. |
| `sort` | string | `recent` | Keyword search order: `recent` (newest first), `relevance` or `oldest`. |
| `includeArticleContent` | boolean | `false` | Also open every found article on Naver News and add full text, exact publish and update minute, sections, image URLs and the original URL. |
| `articleUrls` | array of strings | — | Naver News article addresses to read in full (formats listed under Key features). |
| `sections` | array of strings | — | Any of `politics`, `economy`, `society`, `culture`, `world`, `science`. |
| `pressIds` | array of strings | — | Publisher codes or publisher page addresses, e.g. `001`, `015` or `https://media.naver.com/press/001`. With `ranking` set, returns that publisher's ranking instead of its article list. |
| `ranking` | string | `none` | `none`, `mostViewed` (많이 본 뉴스) or `mostCommented` (댓글 많은 뉴스). |
| `daysBack` | integer | — | 0 to 3,650. Only the last N days in Korea time (0 = today only). For keyword searches and publisher lists. Do not combine with `dateFrom`. |
| `dateFrom` | string | — | `YYYY-MM-DD`, Korea time. If only `dateTo` is set, `dateFrom` is 2000-01-01. |
| `dateTo` | string | — | `YYYY-MM-DD`, Korea time. If only `dateFrom` is set, `dateTo` is today. |
| `onlyNewSinceLastRun` | boolean | `false` | Monitoring mode: return only articles not returned by earlier runs of the same list. |
| `lookbackHours` | integer | `48` | 0 to 720. With `onlyNewSinceLastRun`: how far before the previous complete run each run re-checks for late-listed articles. |
| `historyStoreName` | string | `naver-news-monitor-history` | With `onlyNewSinceLastRun`: the key-value store in your account that remembers returned articles. Letters, digits and hyphens, up to 63 characters. One name per monitor. |
| `proxyConfiguration` | object | Apify Proxy | Keep the default (datacenter). Switch to RESIDENTIAL with country KR only if the log shows repeated request errors. |

Finding a publisher code: open the publisher on Naver News; the address is `media.naver.com/press/<code>`. The code is also the first number in every article address (`/article/<code>/<article number>`) and comes back in every row as `pressId`.

#### How "only new articles since the last run" works

With `onlyNewSinceLastRun` the Actor keeps the IDs of the articles it **actually saved**, one record per list, in `historyStoreName` in your own account. A list is one keyword with its sort and the date range as typed, or one section, one publisher or one ranking, so different searches never share history. Each record keeps the newest 5,000 IDs.

The first run returns the latest articles as usual and becomes the starting point. A later run with `sort: "recent"` reads from the newest article downward, skips known IDs, and stops after 5 articles in a row dated before the previous complete run minus `lookbackHours`. In run `xWkBb863hYlazCIEm` (two keywords, one section and one publisher, started about 40 seconds after the first run of the same input) the Actor made 11 requests, skipped 105 articles it had already returned and saved 10 new ones.

Limits: it is best effort, not a guarantee of complete coverage. An article that Naver lists later than `lookbackHours` after its publish time, or more new articles than `maxResults` between two runs, can be missed. If the first run was cut off by `maxResults`, later runs start from that cut-off and do not go back for older articles. With `sort` other than `recent` the Actor still skips known articles but cannot stop early. Do not run the same monitor twice at the same time. Use `daysBack` rather than fixed dates in a schedule. It does not apply to `articleUrls`. If the history store cannot be opened, the run behaves like a first run and says so in the log.

### What data can you extract from Naver News?

One row per article. Every row has every field; a value that is not known is `null`.

| Field | Type | Description |
|---|---|---|
| `id` | string | Stable key: `naver-news:<pressId>/<articleId>`, or `web:<host><path>` for an article that is only on the publisher site. |
| `foundBy` | string | `search`, `articleUrl`, `section`, `sectionHeadline`, `press`, `rankingMostViewed` or `rankingMostCommented`. |
| `keyword` | string | The keyword that found the row; `null` for rows not found by keyword search. |
| `rank` | integer | Position in the list the row came from. In rankings: the rank Naver shows, within the publisher. `null` for `articleUrls`. |
| `title` | string | Headline. |
| `url` | string | The Naver News link when Naver News hosts the article, otherwise the publisher link. |
| `naverNewsUrl` | string | The article on Naver News; `null` when Naver News does not host it. |
| `originalUrl` | string | The article on the publisher's site. Always present for search rows without a Naver News copy; for other rows it is filled when the article page is read. |
| `pressId`, `articleId` | string | Naver's publisher code and article number. |
| `pressName` | string | Publisher name as Naver shows it. |
| `summary` | string | The snippet or lead Naver shows. `null` in rankings unless the article page is read. |
| `thumbnailUrl` | string | Thumbnail image. |
| `publishedAt` | string | Publish time, ISO 8601 in Korea time (+09:00). |
| `publishedAtPrecision` | string | `minute`, `hour`, `day`, `week` or `month`: how exact `publishedAt` is. |
| `updatedAt` | string | Last update time from the article page; `null` when the article was not updated or the page was not read. |
| `section` | string | Section in Korean (정치, 경제, 사회, 생활/문화, 세계, IT/과학 …): the section the row was found in, or the first section on the article page. |
| `sections` | array of strings | All sections on the article page; `null` when the page was not read. |
| `viewCount`, `commentCount` | integer | Only in one publisher's own ranking (`pressIds` + `ranking`). `null` elsewhere. |
| `content` | string | Full text, one paragraph per line, photo captions removed. `null` unless the article page was read. |
| `contentLength` | integer | Number of characters in `content`. |
| `imageUrls` | array of strings | Images inside the article body. |
| `contentStatus` | string | `null` = article page not requested; `ok`; `empty` (page read but no text, e.g. photo or video only); `unsupported` (no readable copy on Naver News); `notFound`; `failed` (the page did not answer after 3 attempts). |
| `scrapedAt` | string | Collection time (UTC, ISO 8601). |
| `schemaVersion` | integer | Row format version, currently `1`. |

**Run summary** (key-value store record `RUN_SUMMARY`, not charged): `totalReturned`, `chargeLimitReached`, and one entry per list with `kind`, `value`, `status` (`completed`, `partial`, `budgetLimited`, `failed`), `stopReason` (`maxResultsReached`, `endOfResults`, `noResults`, `reachedHistoryWindow`, `chargeLimitReached`, `naverResultLimit`, `repeatedResults`, `requestsFailed`, `layoutChanged`, `invalidInput`), `returnedCount`, `pagesFetched`, `skippedAlreadyReturned`, `skippedBeforeWindow` and `windowStart`. For `articleUrls`: `requested`, `returnedCount`, `duplicatesSkipped` and `notReturned` with a reason per address. Also article-text counts (`ok`, `empty`, `notOnNaver`, `unsupported`, `notFound`, `failed`) and request and timeout counts.

#### Output example: search row with full text

From run `VV3yRjyacYH5eofNg` (2026-11-11, build 0.1.8), input `keywords: ["삼성전자"]`, `sort: "relevance"`, `includeArticleContent: true`. `content` is cut here after its first four lines (the real value has 1,164 characters); everything else is unedited.

```json
{
  "schemaVersion": 1,
  "id": "naver-news:015/0005352683",
  "foundBy": "search",
  "keyword": "삼성전자",
  "rank": 1,
  "title": "\"삼성 거래처도 아닌데 이렇게까지\"…중소기업에 벌어진 일",
  "url": "https://n.news.naver.com/mnews/article/015/0005352683",
  "naverNewsUrl": "https://n.news.naver.com/mnews/article/015/0005352683",
  "originalUrl": "https://www.hankyung.com/article/202611119762g",
  "pressId": "015",
  "articleId": "0005352683",
  "pressName": "한국경제",
  "summary": "삼성전자가 국내 기업 가운데 처음으로 동반성장지수 평가에서 15년 연속 최우수 등급을 받았다. 협력회사에 대한 기술·금융 지원과 인력 양성뿐 아니라 거래 관계가 없는 중소기업까지 지원 대상을 확대한 점을 인정받았다. 삼성전자는 11일 동반성장위원회가 발표한 '2026년도 동반성장지수 평가(2025년...",
  "thumbnailUrl": "https://search.pstatic.net/common/?src=https%3A%2F%2Fimgnews.pstatic.net%2Fimage%2Forigin%2F015%2F2026%2F11%2F11%2F5352683.jpg&type=fface200_200&expire=2&refresh=true",
  "publishedAt": "2026-11-11T12:01:00+09:00",
  "publishedAtPrecision": "minute",
  "updatedAt": "2026-11-11T12:09:00+09:00",
  "section": "경제",
  "sections": ["경제"],
  "viewCount": null,
  "commentCount": null,
  "content": "삼성전자, 동반성장지수 15년 연속 최우수\n스마트공장 3600곳·협력사 27만명 교육\n거래 없는 중소기업까지 기술·자금 지원\n삼성전자가 국내 기업 가운데 처음으로 동반성장지수 평가에서 15년 연속 최우수 등급을 받았다. 협력회사에 대한 기술·금융 지원과 인력 양성뿐 아니라 거래 관계가 없는 중소기업까지 지원 대상을 확대한 점을 인정받았다.\n…",
  "contentLength": 1164,
  "imageUrls": [
    "https://imgnews.pstatic.net/image/015/2026/11/11/0005352683_001_20261111120912893.jpg?type=w860"
  ],
  "contentStatus": "ok",
  "scrapedAt": "2026-11-11T10:07:33.214Z"
}
```

#### Output example: search row without a Naver News copy

From run `yGRypznVbhdUE3aGY` (2026-11-11, build 0.1.7), keyword `환율` with `includeArticleContent: true`. The article is only on the publisher's site, so there is no `pressId`, the time is exact to the hour only, and the text was not read. `summary` is cut here because the snippet contains the reporter's byline; everything else is unedited.

```json
{
  "schemaVersion": 1,
  "id": "web:www.dnews.co.kr/uhtml/view.jsp?idxno=202611111726393660664",
  "foundBy": "search",
  "keyword": "환율",
  "rank": 2,
  "title": "위메이드플레이, 3분기 영업익 21억…환율 하락에 순익 적자",
  "url": "https://www.dnews.co.kr/uhtml/view.jsp?idxno=202611111726393660664",
  "naverNewsUrl": null,
  "originalUrl": "https://www.dnews.co.kr/uhtml/view.jsp?idxno=202611111726393660664",
  "pressId": null,
  "articleId": null,
  "pressName": "대한경제",
  "summary": "매출 303억…전년比 2.6% 감소 환율 하락에 순손실 24억 전환 3매치 신작 내년 초 시험 출시 …",
  "thumbnailUrl": "https://search.pstatic.net/common/?src=https%3A%2F%2Fimgnews.pstatic.net%2Fimage%2Forigin%2F5138%2F2026%2F11%2F11%2F684567.jpg&type=fface200_200&expire=2&refresh=true",
  "publishedAt": "2026-11-11T17:00:00+09:00",
  "publishedAtPrecision": "hour",
  "updatedAt": null,
  "section": null,
  "sections": null,
  "viewCount": null,
  "commentCount": null,
  "content": null,
  "contentLength": null,
  "imageUrls": null,
  "contentStatus": "unsupported",
  "scrapedAt": "2026-11-11T09:57:26.471Z"
}
```

#### Output example: publisher ranking row

From run `nbaSai3VXokdCKXox` (2026-11-15, build 0.1.8), input `pressIds: ["001"]`, `ranking: "mostViewed"`. Unedited. A publisher's ranking page gives no time or snippet, so those fields are `null` unless `includeArticleContent` is on.

```json
{
  "schemaVersion": 1,
  "id": "naver-news:001/0016442797",
  "foundBy": "rankingMostViewed",
  "keyword": null,
  "rank": 4,
  "title": "서울 송파구 잠실광역환승센터 화재…인명피해 없어",
  "url": "https://n.news.naver.com/mnews/article/001/0016442797",
  "naverNewsUrl": "https://n.news.naver.com/mnews/article/001/0016442797",
  "originalUrl": null,
  "pressId": "001",
  "articleId": "0016442797",
  "pressName": "연합뉴스",
  "summary": null,
  "thumbnailUrl": "https://mimgnews.pstatic.net/image/origin/001/2026/11/15/16442797.jpg?type=nf148_148",
  "publishedAt": null,
  "publishedAtPrecision": null,
  "updatedAt": null,
  "section": null,
  "sections": null,
  "viewCount": 5337,
  "commentCount": null,
  "content": null,
  "contentLength": null,
  "imageUrls": null,
  "contentStatus": null,
  "scrapedAt": "2026-11-15T13:57:11.351Z"
}
```

### Advanced usage

#### Send new brand mentions to your own system

Schedule the monitoring input, then read the new rows:

```python
from apify_client import ApifyClient

client = ApifyClient("YOUR_APIFY_TOKEN")
run = client.actor("magenta_courser/naver-news-scraper").call(run_input={
    "keywords": ["올리브영"],
    "onlyNewSinceLastRun": True,
    "historyStoreName": "oliveyoung-news-hourly",
    "includeArticleContent": True,
})
for row in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(row["publishedAt"], row["pressName"], row["title"], row["url"])
```

#### Combine sources in one run

Lists are read three at a time and share one dataset; `foundBy`, `keyword` and `section` tell the rows apart. `maxResults` applies to each list separately.

```json
{
  "keywords": ["코스피"],
  "sections": ["economy"],
  "pressIds": ["015"],
  "maxResults": 30
}
```

(`코스피` = KOSPI; `015` is Korea Economic Daily.)

#### Large full-text runs: raise the memory

Article text is the slow part, one page load per article. The same input (three keywords and one section, 120 rows each, full text) took 297.3 seconds at 256 MB (run `xQqrmIWyQIHM8YgRe`, build 0.1.6) and 113.8 seconds at 1,024 MB (run `yGRypznVbhdUE3aGY`, build 0.1.7). Set the memory in the run options.

#### De-duplicate across runs and sources

The same article can be found by a keyword, a section and a ranking. Use `id` as the key in your own database.

#### One call from the command line

```bash
curl -X POST "https://api.apify.com/v2/acts/magenta_courser~naver-news-scraper/run-sync-get-dataset-items" \
  -H "Authorization: Bearer $APIFY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"keywords": ["삼성전자"], "maxResults": 20}'
```

### How much does it cost to scrape Naver News?

Pay per event, no subscription. Prices below are the Free-plan prices in effect since 2026-11-11; paid Apify plans get the discount shown.

| Event | Price | Bronze / Silver / Gold and above | When |
|---|---|---|---|
| `article` | **$0.0004** per article row | $0.00038 / $0.00036 / $0.00032 | Every row saved to the dataset, with or without full text. |
| Actor start | $0.00005 | — | Once per run. |

Not charged: full article text, the `RUN_SUMMARY` record, list pages that return no rows, articles skipped as already returned in monitoring mode, `articleUrls` addresses that could not be read, lists not started because of your cost limit.

Charged like any other row: a search row that has no copy on Naver News (`contentStatus: "unsupported"`), and a row whose article page did not answer (`contentStatus: "failed"`). Both still carry the list data.

Worked examples at the Free-plan price:

| Job | Events | Cost |
|---|---|---|
| Default run: 100 newest articles for 1 keyword | 100 articles | **$0.04** |
| 1,000 articles with full text | 1,000 articles | **$0.40** |
| 20 keywords, last day, 50 articles each, daily | 1,000 articles per day | **$0.40 per day** |
| Hourly monitor of one keyword that gets 30 new articles a day | 30 articles + 24 starts per day | about **$0.013 per day** |
| One publisher's top 20 most-viewed | 20 articles | **$0.008** |

Set **Maximum cost per run** in the console to cap a run. It stops cleanly: rows already stored are kept, the list being read gets `status: "budgetLimited"`, and lists not started are marked the same way with zero rows. In run `NVymELwvYG0iunf9O` a limit of $0.006 stopped a 20-row request at 14 rows, and 14 `article` events were charged.

### Scheduling and integrations

- **Schedules.** In Apify Console, create an hourly or daily schedule with `onlyNewSinceLastRun: true`. Use one `historyStoreName` per schedule.
- **Webhooks and automation.** Send the dataset to Slack, Google Sheets, Make, Zapier or n8n with Apify integrations, or call your own endpoint with a webhook when a run finishes.
- **API clients.** Any HTTP client works; official clients exist for Python (`apify-client`) and JavaScript (`apify-client`).

```javascript
import { ApifyClient } from 'apify-client';

const client = new ApifyClient({ token: 'YOUR_APIFY_TOKEN' });
const run = await client.actor('magenta_courser/naver-news-scraper').call({
    keywords: ['삼성전자'],
    daysBack: 1,
    includeArticleContent: true,
});
const { items } = await client.dataset(run.defaultDatasetId).listItems();
console.log(items.filter((r) => r.contentStatus === 'ok').length, 'articles with full text');
```

### FAQ

**Do I need a Naver account or the Naver Search API key?**
No. The Actor reads Naver's public news pages without logging in.

**Why do some search rows have no Naver News link?**
Naver's news search also lists articles that live only on the publisher's own site. In run `kkIzYQnrVM6s7RUtH` (`삼성전자`, newest first) 50 of 100 rows had a Naver News copy; in run `yGRypznVbhdUE3aGY` (three keywords by relevance) 302 of 360 did. The others are returned with `naverNewsUrl: null` and their link in `originalUrl`.

**Can it read the full text from the publisher's own site?**
No. Full text is read only from Naver News, Naver Sports and Naver Entertainment article pages. Rows without such a copy get `contentStatus: "unsupported"`.

**Why is `publishedAt` only exact to the hour or day?**
That is what Naver's list shows ("3시간 전", "2일 전"). Turn on `includeArticleContent` for the exact minute. Section headlines and the rows of a publisher's own ranking have no time in the list at all.

**How many articles can one keyword return?**
`maxResults` goes up to 1,000, and Naver itself lists about 1,000 results per search. To go further back, split the period with `dateFrom` / `dateTo` and run once per range.

**What happens when a page is slow or blocked?**
List pages are tried up to 5 times, each time on a new proxy session. Article pages are tried 3 times with a 12 to 15 second timeout; if all attempts fail the row is still delivered with `contentStatus: "failed"`, and the count is in `RUN_SUMMARY`. In run `xQqrmIWyQIHM8YgRe` 19 of 438 article requests timed out and were repeated, and no article ended as `failed`. If Naver changes a page layout, the affected list is marked `failed` with `stopReason: "layoutChanged"`, the run ends as failed, and the articles that could be read are still saved.

**Does it collect reporter names or comments?**
Not as fields. Reporter names and profiles, comments, comment writers and reactions are not read. The article text is delivered as the publisher wrote it, so a byline line inside the text (for example "(서울=연합뉴스) … 기자 =") stays part of `content`, and a snippet can contain one too. `commentCount` is a number only.

**Is it legal to scrape Naver News?**
The Actor collects only publicly visible pages and does not log in. Article text belongs to its publisher: check the publisher's terms before you republish or redistribute it. Whether your use is allowed depends on Naver's terms of service, your purpose and your local law; checking that is your responsibility.

**네이버 뉴스 기사를 엑셀로 받을 수 있나요?**
네. 실행이 끝나면 데이터셋을 JSON·CSV·Excel로 내려받을 수 있습니다. 본문까지 필요하면 `includeArticleContent`를 `true`로 두면 되고, 추가 요금은 없습니다.

### More Korean data Actors

[![Naver AI Briefing Scraper: $0.002 per AI Briefing found](https://raw.githubusercontent.com/experimory-crypto/naver-map-scraper-examples/main/assets/hero-naver-ai-briefing-scraper.png)](https://apify.com/magenta_courser/naver-ai-briefing-scraper)

[![Naver Shopping Rank Tracker: top 1,000 ranks for any keyword](https://raw.githubusercontent.com/experimory-crypto/naver-map-scraper-examples/main/assets/hero-naver-shopping-rank-tracker.png)](https://apify.com/magenta_courser/naver-shopping-rank-tracker)

| Actor | What you put in → what you get |
|---|---|
| [Naver Search Scraper](https://apify.com/magenta_courser/naver-search-scraper) | Keywords → the Naver integrated search page as rows: section, rank, ad flag, title, URL, snippet, related searches. |
| [Naver DataLab Search Trends Scraper](https://apify.com/magenta_courser/naver-datalab-trends-scraper) | Korean keywords → search interest (0 to 100) per day, week or month. |
| [Naver AI Briefing Scraper](https://apify.com/magenta_courser/naver-ai-briefing-scraper) | Search queries → Naver's AI answer, its cited sources, and whether your domain is cited. |
| [Naver Blog, Cafe & News Search Scraper](https://apify.com/magenta_courser/naver-blog-cafe-search-scraper) | Keywords → blog, cafe and news search results with title, URL, snippet and publish time. |
| [Naver Shopping Scraper & Keyword Rank Tracker](https://apify.com/magenta_courser/naver-shopping-rank-tracker) | Shopping keywords → ranked products with price, store, rating, reviews and stock. |
| [Naver Map Scraper](https://apify.com/magenta_courser/naver-place-search-scraper) | Keyword → businesses with address, phone, coordinates, rating, review counts. |
| [Naver Map Reviews Scraper](https://apify.com/magenta_courser/naver-place-reviews-scraper) | Place IDs or Naver Map URLs → visitor reviews with rating, text, dates, keyword codes and owner reply. |
| [Naver Place Multi-Location Review Monitor](https://apify.com/magenta_courser/multi-location-review-monitor) | Your locations → new reviews, owner replies added or edited, unanswered counts per location. |
| [Naver Place Rank Tracker](https://apify.com/magenta_courser/naver-place-rank-tracker) | Keywords + your place IDs → your rank in Naver Map search, with change since the last run. |
| [Kakao Map Reviews Scraper](https://apify.com/magenta_courser/kakaomap-reviews-scraper) | Kakao Map place IDs → visitor reviews, no reviewer personal data. |
| [Kakao Map Scraper](https://apify.com/magenta_courser/kakaomap-place-search-scraper) | Keyword → Kakao Map businesses, same field names as the Naver Map Actor. |
| [Korea MCP Server: Naver, Kakao Map](https://apify.com/magenta_courser/korea-local-data-mcp) | An MCP server with Naver Map search, Naver Place reviews and Kakao Map search as tools. |

Code examples for these Actors: [github.com/experimory-crypto/naver-map-scraper-examples](https://github.com/experimory-crypto/naver-map-scraper-examples).

### 한국어 안내 (Korean)

검색어, 기사 URL, 섹션, 언론사 코드, 랭킹 중 하나 이상을 넣으면 네이버 뉴스 기사가 한 건당 한 행으로 나옵니다. 제목, 언론사, 발행 시각과 수정 시각, 네이버 뉴스 URL, 언론사 원문 URL, 썸네일, 섹션이 들어 있고, `includeArticleContent`를 켜면 본문 전문과 본문 이미지 URL이 같은 행에 추가됩니다. 언론사별 랭킹에서는 조회수 또는 댓글 수도 나옵니다. 요금은 기사 행 1건당 $0.0004(1,000건 $0.40)이고 본문 전문에는 추가 요금이 없으며, 실행마다 시작 요금 $0.00005가 붙습니다. "지난 실행 이후 새 기사만" 모드는 이미 받은 기사를 기억해 새 기사만 돌려줍니다(완전한 수집을 보장하지는 않습니다). 제공하지 않는 것: 네이버 뉴스에 사본이 없는 기사(언론사 사이트에만 있는 기사)의 본문, 댓글 내용, 검색 한 번당 약 1,000건을 넘는 결과. 기자 이름·프로필, 댓글과 댓글 작성자는 항목으로 수집하지 않고 이메일 주소는 가려지지만, 본문 안에 적힌 바이라인은 그대로 남습니다. 로그인 없이 공개된 페이지만 읽습니다.

### 中文说明 (Chinese)

输入关键词、文章 URL、版块、媒体代码或排行榜中的至少一项，即可获得 Naver 新闻（네이버 뉴스）文章，每篇文章一行。包含标题、媒体名称、发布时间和更新时间、Naver 新闻 URL、媒体原文 URL、缩略图和版块；开启 `includeArticleContent` 后，同一行还会加入正文全文和正文图片 URL。在单个媒体的排行榜中还会返回浏览量或评论数。费用为每行文章 $0.0004（1,000 篇 $0.40），正文全文不另收费，每次运行另收 $0.00005 启动费。“仅上次运行之后的新文章”模式会记住已返回的文章，只返回新文章（不保证完全覆盖）。不提供：在 Naver 新闻上没有副本的文章（只存在于媒体网站的文章）的正文、评论内容，以及每次搜索约 1,000 条以外的结果。不把记者姓名和个人资料、评论及评论者作为字段采集，电子邮件地址会被遮盖，但正文中写明的署名行会保留。无需登录，只读取公开页面。

### 日本語ガイド (Japanese)

キーワード、記事 URL、セクション、媒体コード、ランキングのうち 1 つ以上を入力すると、Naver ニュース（네이버 뉴스）の記事が 1 件 1 行で出力されます。タイトル、媒体名、公開日時と更新日時、Naver ニュースの URL、媒体の元記事 URL、サムネイル、セクションが含まれ、`includeArticleContent` をオンにすると本文全文と本文内の画像 URL が同じ行に追加されます。媒体ごとのランキングでは閲覧数またはコメント数も取得できます。料金は記事 1 行あたり $0.0004（1,000 件で $0.40）で、本文全文に追加料金はかかりません。実行ごとに開始料金 $0.00005 がかかります。「前回実行以降の新着記事のみ」モードは、すでに返した記事を記憶して新しい記事だけを返します（完全な網羅を保証するものではありません）。提供しないもの：Naver ニュースにコピーがない記事（媒体サイトにのみある記事）の本文、コメントの内容、1 回の検索で約 1,000 件を超える結果。記者名やプロフィール、コメントとその投稿者は項目として収集せず、メールアドレスは伏せられますが、本文中に書かれた署名行はそのまま残ります。ログインなしで公開ページのみを読み取ります。

### Notes

- **Times.** `publishedAt` and `updatedAt` are ISO 8601 in Korea time (+09:00); `scrapedAt` is UTC. A relative label such as "3시간 전" is converted using the run time and truncated to the stated precision.
- **Sports and entertainment articles** live on Naver Sports and Naver Entertainment. They are read too; `naverNewsUrl` keeps the `n.news.naver.com` form.
- **Section feeds** are what Naver shows on the section page: headlines first (no time in the list), then the article feed. The feed is Naver's own selection of recent articles; for a strict newest-first list of one publisher use `pressIds`.
- **Rankings** are per publisher; Naver has no single all-publisher top list. The all-publisher page shows about 5 articles per publisher without counts; counts come only from one publisher's own ranking.
- **Dates** (`daysBack`, `dateFrom`, `dateTo`) apply to keyword searches and publisher lists, not to sections or rankings.
- **Article pages are read from `news.naver.com/article/…`**, with a new proxy session per attempt. Rows are still stored in list order.
- **Article text belongs to its publisher.** E-mail addresses are replaced with `[email redacted]`; the rest of the text is delivered as published.
- Not affiliated with Naver.

### Support

Missing a field, another input mode, a different output format? Tell us in the **Issues** tab with the run ID. We answer every issue.

# Actor input Schema

## `keywords` (type: `array`):

Search keywords, one per line, exactly as you would type them into Naver News search. Korean keywords return the most results, e.g. 삼성전자 (Samsung Electronics), 반도체 수출 (chip exports), or a brand or person name. Each keyword is searched separately. Required unless you fill Article URLs, Sections, Publisher codes or Ranking instead.

## `maxResults` (type: `integer`):

Maximum rows for each keyword (1-1000). Default 100. The same limit applies to each section, each publisher and each ranking. Naver lists about 1,000 results per search, so to go further back split the period with dateFrom/dateTo and run once per period.

## `sort` (type: `string`):

Order of keyword search results: recent (newest first, default; required for monitoring to stop at the previous run), relevance (Naver's relevance ranking) or oldest (oldest first).

## `includeArticleContent` (type: `boolean`):

false (default) = list data only: title, publisher, snippet, time, links. true = also open every found article on Naver News and add its full text (content), exact publish and update minute, section(s), image URLs and the publisher's original URL. Only articles that have a copy on Naver News can be opened; the others keep their row with contentStatus unsupported. One extra page load per article, so runs take longer.

## `articleUrls` (type: `array`):

Optional. Naver News article addresses to read in full, one per line, no keyword needed. Accepts https://n.news.naver.com/article/<publisher>/<article>, .../mnews/article/..., m.news.naver.com and news.naver.com links, Naver Sports and Naver Entertainment article links, old read.naver?oid=..\&aid=.. links, or just 001/0016436704. Each readable article becomes one row with full text; addresses that cannot be read are listed with the reason in RUN_SUMMARY and are not charged.

## `sections` (type: `array`):

Optional. Naver News sections to read: the section's current headlines followed by its article feed (newest first as Naver lists it). Any of: politics (정치), economy (경제), society (사회), culture (생활/문화), world (세계), science (IT/과학).

## `pressIds` (type: `array`):

Optional. Naver News publisher (press) codes or publisher page addresses, one per line, e.g. 001 (Yonhap News), 015 (Korea Economic Daily) or https://media.naver.com/press/001. Returns that publisher's articles on Naver News newest first, day by day (the last 7 days unless you set dates). If Ranking is also set, returns that publisher's ranking instead (up to 20 articles with view or comment counts).

## `ranking` (type: `string`):

Optional. none (default), mostViewed (많이 본 뉴스) or mostCommented (댓글 많은 뉴스). Without Publisher codes: Naver's ranking page with the top 5 articles of about 80 publishers (about 400 rows, so raise Max results to 500 to get all). With Publisher codes: each publisher's top 20 with the view count or comment count Naver shows.

## `daysBack` (type: `integer`):

Optional. Only articles from the last N days, counted in Korea time: 0 = today only, 1 = yesterday and today, 7 = the last week. Stays correct in a schedule because it moves with the run date. Applies to keyword searches and publisher lists. Do not combine with dateFrom.

## `dateFrom` (type: `string`):

Optional. Only articles published on or after this date, Korea time, format YYYY-MM-DD, e.g. 2026-09-01. Applies to keyword searches and publisher lists. If only dateTo is set, dateFrom defaults to 2000-01-01.

## `dateTo` (type: `string`):

Optional. Only articles published on or before this date, Korea time, format YYYY-MM-DD, e.g. 2026-09-15. If only dateFrom is set, dateTo defaults to today.

## `onlyNewSinceLastRun` (type: `boolean`):

Monitoring mode. true = return only articles this Actor has not returned before for the same list (same keyword, sort and date range, or the same section, publisher or ranking; remembered in the history store below). The first run returns the latest articles as usual. Use with sort recent and a schedule (e.g. hourly or daily). Best effort, not a guarantee of complete coverage: articles that Naver lists later than lookbackHours after their publish time, or more new articles than Max results between runs, can be missed. Default false.

## `lookbackHours` (type: `integer`):

Only used when onlyNewSinceLastRun is true. Each run re-checks articles dated up to this many hours before the previous complete run started, so articles that appear in Naver late are still returned once. Larger = fewer missed late articles, a few more pages read (pages are not charged, only returned rows). 0-720, default 48.

## `historyStoreName` (type: `string`):

Only used when onlyNewSinceLastRun is true. Name of the key-value store in your Apify account that remembers already returned articles. Inside it every list has its own record, so different searches never share history. Use a separate name for each separate monitor, e.g. brand-news-hourly, and do not run the same monitor twice at the same time (it can return some articles twice). Letters, digits and hyphens only, max 63 characters.

## `proxyConfiguration` (type: `object`):

Proxy settings. The default Apify datacenter proxy works for Naver News. Switch to RESIDENTIAL with country KR only if the log shows repeated request errors.

## Actor input object example

```json
{
  "keywords": [
    "삼성전자"
  ],
  "maxResults": 100,
  "sort": "recent",
  "includeArticleContent": false,
  "ranking": "none",
  "onlyNewSinceLastRun": false,
  "lookbackHours": 48,
  "historyStoreName": "naver-news-monitor-history",
  "proxyConfiguration": {
    "useApifyProxy": true
  }
}
```

# Actor output Schema

## `articles` (type: `string`):

All scraped news articles as dataset items, one row per article.

## `runSummary` (type: `string`):

Free (not charged). One entry per list (keyword search, section, publisher or ranking): status (completed / partial / budgetLimited / failed), returnedCount, stopReason, pages read, articles skipped as already returned and the monitoring window used. For Article URLs: the addresses that were not returned with the reason for each. Also article text counts (ok, not on Naver News, failed) and request / timeout counts.

# API

You can run this Actor programmatically using our API. Below are code examples in JavaScript, Python, and CLI, as well as the OpenAPI specification and MCP server setup.

## JavaScript example

```javascript
import { ApifyClient } from 'apify-client';

// Initialize the ApifyClient with your Apify API token
// Replace the '<YOUR_API_TOKEN>' with your token
const client = new ApifyClient({
    token: '<YOUR_API_TOKEN>',
});

// Prepare Actor input
const input = {
    "keywords": [
        "삼성전자"
    ],
    "proxyConfiguration": {
        "useApifyProxy": true
    }
};

// Run the Actor and wait for it to finish
const run = await client.actor("magenta_courser/naver-news-scraper").call(input);

// Fetch and print Actor results from the run's dataset (if any)
console.log('Results from dataset');
console.log(`💾 Check your data here: https://console.apify.com/storage/datasets/${run.defaultDatasetId}`);
const { items } = await client.dataset(run.defaultDatasetId).listItems();
items.forEach((item) => {
    console.dir(item);
});

// 📚 Want to learn more 📖? Go to → https://docs.apify.com/api/client/js/docs

```

## Python example

```python
from apify_client import ApifyClient

# Initialize the ApifyClient with your Apify API token
# Replace '<YOUR_API_TOKEN>' with your token.
client = ApifyClient("<YOUR_API_TOKEN>")

# Prepare the Actor input
run_input = {
    "keywords": ["삼성전자"],
    "proxyConfiguration": { "useApifyProxy": True },
}

# Run the Actor and wait for it to finish
run = client.actor("magenta_courser/naver-news-scraper").call(run_input=run_input)

# Fetch and print Actor results from the run's dataset (if there are any)
print(f"💾 Check your data here: https://console.apify.com/storage/datasets/{run.default_dataset_id}")
for item in client.dataset(run.default_dataset_id).iterate_items():
    print(item)

# 📚 Want to learn more 📖? Go to → https://docs.apify.com/api/client/python/docs/quick-start

```

## CLI example

```bash
echo '{
  "keywords": [
    "삼성전자"
  ],
  "proxyConfiguration": {
    "useApifyProxy": true
  }
}' |
apify call magenta_courser/naver-news-scraper --silent --output-dataset

```

## MCP server setup

```json
{
    "mcpServers": {
        "apify": {
            "type": "http",
            "url": "https://mcp.apify.com/?tools=fetch-actor-details,magenta_courser/naver-news-scraper"
        }
    }
}
```

The hosted server signs you in with OAuth on first connect, so no API token belongs in this config. Clients without OAuth support can send an `Authorization: Bearer <APIFY_API_TOKEN>` header instead, using a token from API & Integrations in Apify Console (https://console.apify.com/settings/integrations).

## OpenAPI specification

Download the OpenAPI definition: https://api.apify.com/v2/actors/cpN1GPyl1f65kfEvT/builds/j3hV0bRe4CRihNjY3/openapi.json
