Naver SEO tracking
Check where your pages, blog posts and cafe threads appear on Naver for each target keyword. Store position, link and result type per day to see what moves.
- 1search
- Rank history
Naver search results, typed: title, link, snippet, source and position for any Korean or English query, across the integrated, web, news, image and video tabs. Public data only.
One endpoint covers Naver search. Send one keyword, pick the tab, and get clean JSON back for 2 credits per successful call.
Pick an endpoint, fill in the input and press Run. Without a key you see the catalog sample and nothing is billed. With your API key it runs a real call from your browser.
Every successful call costs 1 credit, whether the page holds one profile or a page of comments. Failed calls cost nothing. A free tier lets you try it first.
On Pro, 1,000 calls at 1 credit cost $1.58. Scale brings that to $1.20, so 250,000 Naver calls a month is $299.
See all plans →| Plan | Price / mo | Credits / mo | Per 1,000 calls |
|---|---|---|---|
| Free | $0 | Free tier | — |
| Hobby | $29 | 15,000 | $1.93 |
| Pro | $79 | 50,000 | $1.58 |
| Scale | $299 | 250,000 | $1.20 |
Built to be easy to start with, even if you've never scraped a search engine. Here's how to get Naver results in a few minutes:
naver-search returns the results for one query on search.naver.com.queries per call, plus where: nexearch, web, news, image or video.next_cursor; pass it back as cursor for the next page. null means the last page. The tab stays locked in the cursor.X-API-Key header, or ask your AI agent over MCP.Every endpoint takes the same kind of input.
You don't know the items yet: find them with a keyword.
서울 맛집// GET /api/v2/naver-search/ { "queries": "서울 맛집", "where": "nexearch" }
| Send this | Parameter | And you can get |
|---|---|---|
| Keyword | queries | Results with title, link, snippet, source and position |
| Search tab | where | Integrated (nexearch), web, news, image or video results |
| Cursor | cursor | The next page of the same query and tab |
| Raw mode | raw_data | The raw upstream payload instead of the formatted results |
Optional on list endpoints: cursor (pass back next_cursor for the next page). Auth is one header: X-API-Key. Send parameters in the query string (GET) or as a JSON body (POST).
The one thing to know: each data type has its own shape. Field names stay the same on every call, so code and agents read them directly. Samples below come from the live catalog.
// GET /api/v2/naver-search/?queries=서울 맛집 // one item of data.results · catalog sample { "result_type": "web_organic", "query": "서울 맛집", "where": "nexearch", "position": 1, "title": "서울 맛집좀 잘 아시는분? ㅋ", "link": "https://cafe.naver.com/divclub/60843", "snippet": "서울에 맛집좀 잘 아시는분 있을까요? ㅋ 대단한집 아니여도 오래된 식당도 좋고...", "source": "배당 투자자 모임 [미국주식,ETF,장기투자,은퇴,파이어족]새 창 열림", "fetched_at": "2026-09-16T07:13:15.311447+00:00" }
Every response is {"success": true, "data": {…}}: your input echoed back, the result under its own key (profile, posts, comments…) and next_cursor for the next page.
Public Naver results power SEO and marketing in South Korea, news monitoring and market research. Here are three common jobs and the Naver scraper endpoint each one uses.
Check where your pages, blog posts and cafe threads appear on Naver for each target keyword. Store position, link and result type per day to see what moves.
Search your brand, product or competitor names with where set to news. Collect new headlines and sources so your team sees coverage in Korean media as it appears.
Read what ranks on Naver for a category before you launch there: which blogs, cafes and shops show up and what they say. Compare tabs to see where your audience looks.
Naver Scraper fetches clean, structured Naver search results at a fixed cost per call: built for Korean SEO, news monitoring and AI agents.
Each request reads Naver at call time, and every result says when in fetched_at.
Same field names on every call, numbers as numbers, ready for code, sheets or an LLM.
Results come a page at a time with next_cursor; run up to 3 requests in parallel per key.
Integrated, web, news, image and video results share the same fields and envelope.
The Scrapier MCP server reaches every Naver endpoint from Claude Code, Cursor, Windsurf and VS Code with one command.
Plain REST: Python, Node.js, cURL, n8n, Make, Zapier or Google Sheets.
Public data only. No Naver account, developer app or client secret; Scrapier routes the request for you.
When Naver changes its pages, Scrapier updates the scraper; your field names stay the same.
2 credits per call, however many results the page holds. Failed calls are never billed.
The official Naver Search API needs a Naver Developers app with a client ID and secret and has its own daily call quota. Scrapier reads the public Naver result pages with one API key, at a price you know before the call runs.
Any non-2xx response is free, so retries cost nothing.
At 2 credits a call. Flat per-call pricing, no compute units or bandwidth to estimate.
No Naver Developers app, client ID or client secret to register.
The same key and JSON envelope cover Google, Amazon, TikTok, YouTube and more.
| Feature | Scrapier Naver Scraper | Naver Search API | DIY scraper |
|---|---|---|---|
| Whose data | Public Naver result pages | Naver search results, per API category | Public Naver result pages |
| Setup | One API key | Naver Developers app, client ID and secret | Proxies, browsers, parsers |
| Maintenance | Handled by Scrapier | App and quota management | Breaks on layout changes |
| Cost | From $1.20 per 1,000 calls | Free within a daily quota | Servers, proxies and engineering time |
| AI agents | MCP server included | Build your own | Build your own |
Naver results for one query, under results. Each result has result_type, position, title, link, snippet, source, the tab in where, the query you sent and fetched_at. The response also gives next_cursor for the next page.
Set where to nexearch for the integrated results page, or to web, news, image or video for one tab. The default is nexearch. The tab is locked into the cursor, so every page of one query stays on the same tab.
Yes. Send the query exactly as a Korean user would type it, for example 서울 맛집. English queries work too. Each result repeats the query, so you can match results to keywords in your own tables.
Read next_cursor from the response and send it back as cursor. Each page is one call at 2 credits. When next_cursor is null you have reached the last page Naver shows for that query.
No. Each call takes one keyword in queries. To track a list, loop over it and run up to 3 calls in parallel per key; failed or throttled calls cost nothing, so retries are free.
No. Scrapier reads public Naver result pages for you, so there is no Naver account, developer app or client secret to create. You only need a Scrapier API key in the X-API-Key header.
Naver Search costs 2 credits per successful call, however many results the page holds. Failed calls are free. Plans: Hobby $29 for 15,000 credits, Pro $79 for 50,000, Scale $299 for 250,000, Enterprise custom. A free tier lets you try it first.
Yes. Set raw_data to true and the response carries the raw upstream payload under raw_responses instead of the formatted results. Pagination with cursor and next_cursor works the same way in both modes.
It depends on what you collect, where you are and how you use it, so this is not legal advice. Scrapier returns publicly visible Naver data only. Some of it can be personal data under GDPR or CCPA, so collect only what you have a legitimate reason to use and check Naver's terms.
Yes. Any tool that can make an HTTP request works: n8n, Make, Zapier, Google Sheets, Airbyte or your own code. Send your key in the X-API-Key header and read the JSON under data.
Yes. Run claude mcp add -t http scrapier https://mcp.scrapier.io/mcp with your key in an X-API-Key header. The agent finds the right Naver endpoint with search_endpoints and calls it with run_scrape, in Claude Code, Cursor or Windsurf.
Check the status code first: 401 means the X-API-Key header is missing, 403 an invalid key, 402 no credits left, 429 too many requests (wait Retry-After seconds). If it's still wrong, report it with the request and the response.
Run up to 3 requests in parallel per API key. If you go faster, the API answers 429 with a Retry-After header: wait that many seconds and retry. 429 and 5xx responses are never billed, and long-running endpoints need a client timeout of 210 seconds or more.
Use the requests library: requests.get("https://proxy.scrapier.io/api/v2/naver-search/", headers={"X-API-Key": key}, params={"queries": "서울 맛집"}, timeout=210). The response is JSON, so read r.json()["data"]["results"] directly; there is no HTML to parse. Each call returns one page.
Free API key. No sales call, no proxy setup, and failed calls are free.