Naver Open Search API — Web, Blog, News, Cafe, Image

Query Naver Search through the official NAVER API HUB and get one clean dataset row per result. Built for Korean SEO teams, brand monitors, and AI agents that need structured blog, news, web, cafe, image, Knowledge iN, encyclopedia, or local results without browser scraping. Runs on Apify with API, schedule, and MCP access.

Best fit

This Actor is designed for teams that already hold NAVER API HUB credentials and want official Search JSON for Korean keywords. It works well for brand monitoring, content research, and agent tool calls that need a bounded, BYOK Open API path.

Constructive routing: start with a small maxResults sample here, then continue with Naver Product Shopping Scraper when you need shopping product cards, or pair it with Naver Map Scraper for place phones and hours. For multi-country Google SERP blocks, use Google SERP All-in-One Scraper.

Supported verticals: blog, news, web, cafe, image, kin, encyclopedia, local, plus adult and errata helpers. Shop, book, and document Search endpoints are outside this Actor's contract because Naver discontinued them without a HUB replacement — choose the Shopping sibling for product catalogs instead.

Practical scenario

An SEO lead enters 강남 맛집 with searchType: "blog" and their NAVER API HUB Client ID/Secret. The run returns ten blog rows with title, link, snippet, blogger name, and post date. They export CSV, then schedule the same input weekly for content monitoring.

Quick start input

{
  "query": "강남 맛집",
  "searchType": "blog",
  "maxResults": 10,
  "authStyle": "hub",
  "naverClientId": "YOUR_CLIENT_ID",
  "naverClientSecret": "YOUR_CLIENT_SECRET"
}

Create an application in the NAVER Cloud Platform console under NAVER API HUB, enable Search, and paste Client ID / Client Secret into the secret fields. Prefer authStyle: "hub" for new applications. Bring your own keys for every run.

Input reference

Field	Type	What it controls
query / queries	string / array	Keywords to search (up to 50 unique).
searchType	enum	blog, news, web, cafe, image, kin, encyc, local, adult, errata.
maxResults	integer	Run-wide row cap (default 30, max 1000).
maxResultsPerQuery	integer	Per-query pagination budget.
sort	enum	sim/date for most lists; random/comment for local.
authStyle	enum	hub (recommended) or legacy.
naverClientId / naverClientSecret	secret	Caller must supply their own BYOK credentials (isSecret; never written to dataset, output, or logs).

What data you receive

One dataset item is one Naver search result (or one adult/errata helper row).

{
  "query": "강남 맛집",
  "searchType": "blog",
  "rank": 1,
  "title": "Best restaurants in Gangnam",
  "link": "https://blog.naver.com/example/123",
  "description": "A short snippet from Naver Search.",
  "bloggerName": "foodie",
  "postDate": "20260917",
  "source": "naver-api-hub",
  "collectedAt": "2026-09-17T12:00:00.000Z"
}

Field	Meaning
rank	1-based position within that query.
title / description	Cleaned text (HTML highlights stripped by default).
source	naver-api-hub or naver-openapi-legacy.

OUTPUT and RUN_SUMMARY store outcome, itemsPushed, and chargedEventCounts. Download JSON, CSV, Excel, or HTML from the dataset.

Pricing

This Actor uses Pay per event plus platform usage pass-through. Always check the live Pricing tab on the Actor page for the current event map — Treat the live Pricing tab as the current source of truth for event prices.

Event	Price
apify-actor-start	$0.00005 / run
search-result	$0.005 / persisted row

Upstream Naver Search is temporarily free within NAVER API HUB quotas when you bring your own keys. You pay Apify events plus platform compute on each run.

Typical cost: 10 blog rows ≈ $0.05005 in events (plus small platform usage).

Use through the API

curl -X POST "https://api.apify.com/v2/acts/khadinakbar~search-naver-open-api/runs" \
  -H "Authorization: Bearer $APIFY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"query":"강남 맛집","searchType":"blog","maxResults":5,"naverClientId":"...","naverClientSecret":"..."}'

Agent / MCP prompt

Use khadinakbar/search-naver-open-api (MCP tool apify--search-naver-open-api) when an AI agent needs official Naver Search JSON. Pass query, searchType, and caller-supplied naverClientId/naverClientSecret. Read OUTPUT.outcome and the default dataset for provenance (source, collectedAt). Bound cost with maxResults / pagination. Scope is HUB Search verticals only; for shopping or map places, call the sibling Actors named above.

Best results

Provide Korean keywords, start with maxResults: 10 to validate credentials, confirm Search is enabled on your HUB application, and schedule focused weekly monitors rather than unbounded dumps. Local search is capped at five public results per query by Naver — keep that specific bound in mind when you plan coverage.

Builder's note

I built this as an HTTP-only Open API wrapper after sibling Naver Shopping and Map Actors needed Korean residential browsers for HTML. I found that NAVER API HUB Search stays cookieless and predictable for agents when callers bring their own Client ID/Secret, and my goal was a focused standalone Search workflow that pairs cleanly with those deeper scrapers.

FAQ and responsible use

Use only credentials you are authorized to hold, and only for lawful collection of public search metadata under applicable laws and Naver API terms. Naver and NAVER API HUB are trademarks of Naver Corp. / NAVER Cloud. This independent Actor is not affiliated with, associated with, or endorsed by Naver. Report issues on the Actor Issues tab.