Naver Map Scraper - Korean places, reviews, phone numbers and GPS

Scrape Naver Map search results for South Korea and get the business name, category, road address, phone number, visitor review score, review counts and GPS coordinates as clean rows. This Naver Map scraper reads the same place list Korean users see, which is the only complete local business directory for the country, because Google Maps does not have one.

What you get

• 50 places per page, paged automatically until your row limit
• Naver place id and a direct link to the place page on Naver Map
• Business name, category line and Naver's numeric category codes, which survive a category being renamed
• Road-name address, lot-number address and the city / ward / neighbourhood line, so you can group results by area without parsing strings
• Public phone number, falling back to Naver's forwarding number when that is all the business publishes
• Visitor review score out of 5, visitor review count, blog and cafe review count and the combined total
• Opening hours line, owner blurb and the short quoted highlights Naver shows on the card
• Latitude and longitude on every row, already the right way round
• Naver Booking and Naver Pay flags, a newly-opened flag, and the service tags Naver lists such as parking or takeaway
• A run summary that separates "this search found nothing" from "the site did not serve us", because in a dataset those two look identical and mean completely different things

Why scrape Naver Map

South Korea is the one large market where Google Maps is not the local business directory. Korean export restrictions on mapping data mean Google ships a degraded product there, so Koreans search for restaurants, clinics, salons and hotels on Naver Map instead. If you want Korean local business data, Naver is not one source among several, it is the source.

The Naver Local Search open API exists but returns five results per call with a handful of fields and no review data at all, which is not enough to build anything on. The place list behind Naver Map carries the review scores, the review counts, the coordinates and the booking flags, and those are exactly the fields that make local business data worth collecting.

Review counts are the part people underestimate. Naver splits them in two: visitor reviews written by people who were there, and blog and cafe posts that Naver has linked to the business. Those two numbers move for different reasons. Visitor reviews track actual footfall; blog counts track marketing spend, because sponsored blog posting is a normal part of Korean local marketing. Having both on the same row is how you tell a busy restaurant from a well-promoted one.

Input

Field	Type	Required	What it does
query	string	yes	What to search for, for example 강남 카페 or 서울 치과
businessType	select	no	Which Naver list to search: all places, restaurants, clinics, stays or attractions
latitude	string	no	Centre the search on a point, pair with longitude
longitude	string	no	Centre the search on a point, pair with latitude
maxItems	integer	no	Stop after this many places, default 50

Korean queries give the fullest results, which is a property of the index rather than of this actor: Naver files businesses under Korean names, so a Korean query matches more of them. Latin-script brand names work where the business itself uses one.

Output

{
  "place_id": "1668757640",
  "name": "퍼스트커피랩 강남점",
  "category": "카페,디저트",
  "category_codes": ["220036", "220052", "220563"],
  "business_type": "restaurant",
  "url": "https://map.naver.com/p/entry/place/1668757640",
  "road_address": "봉은사로6길 39 1층",
  "address": "서울특별시 강남구 역삼동 …",
  "district": "서울 강남구 역삼동",
  "phone": "0507-1315-7680",
  "visitor_review_score": 4.91,
  "visitor_review_count": 3177,
  "blog_review_count": 2078,
  "total_review_count": 5255,
  "description": null,
  "micro_review": [],
  "image_url": "https://ldb-phinf.pstatic.net/…",
  "image_count": 12,
  "has_booking": true,
  "has_npay": false,
  "is_new": false,
  "options": [],
  "latitude": 37.5026235,
  "longitude": 127.0279481,
  "found": true,
  "blocked": false
}

Review counts arrive from Naver as strings with thousands separators. They leave this actor as numbers, so you can sort and average them without cleaning first.

Use cases

Lead lists for anyone selling to Korean small business. A POS vendor, a delivery platform, a supplier or an agency all need the same thing: every business of a given type in a given area, with a phone number and an address that matches the one on their paperwork. One query per district, the road address and the phone on every row, and the review counts as a rough size proxy, gives you a list you can work through rather than a map you have to click.

Local SEO and listing audits. Korean local search is a Naver problem, not a Google problem. Pulling every competitor in a category and neighbourhood gives you the ranking order Naver actually shows, along with each competitor's review score, visitor review count and blog review count. Tracking that weekly tells you whether a client is gaining on the businesses above them, and the split between visitor and blog counts tells you whether a rival is earning attention or buying it.

Market entry and site selection. Before opening anything in Korea you need to know what is already there. Search a category across a list of neighbourhoods and you get density, the review-score distribution and the coordinates to plot it. A district with forty cafes averaging 4.2 is a different proposition from one with twelve averaging 4.8, and both are different from one where half the entries are flagged as newly opened.

Travel and hospitality data. The stays and attractions surfaces return hotels, guesthouses, museums and parks with the same field set as restaurants. Combined with coordinates and Naver Booking flags, that is enough to build an itinerary tool or to enrich an existing travel product with the Korean listings it is missing.

How it compares

	This actor	delicious_zebu/naver-map-search-results-scraper	huggable_quote/naver-map-scraper
Price per place	$0.0025	$0.0015 plus $0.00005 per run	$0.003 plus $0.00005 per run
Cost of a 200-row run	$0.50	$0.30	$0.60
Both review counts split out	yes	not published	not published
Coordinates on every row	yes	not published	yes
Category codes as well as names	yes	no	no
Says when it was refused	yes	no	no
Reviews text	no	no	yes

delicious_zebu is the cheaper option and the honest comparison says so. What you get here for the difference is the visitor / blog review split, Naver's numeric category codes, and a run that tells you when it was refused instead of handing you an empty dataset. huggable_quote reads individual review text, which this actor does not do at all; if you need the reviews themselves rather than the counts, that is the one to use.

Pricing

Billing is pay-per-event at $0.0025 per place delivered. A refused read costs nothing: if the site does not serve us, you get a row saying so and no charge for it. All pricing is pay-per-event — you only pay for results you receive. No actor-start fee, no per-compute-unit charges.

Limits and gotchas

• Naver Map covers South Korea. Addresses, categories and business names come back in Korean, and that is the data, not a setting.
• One page holds 50 places, so a maxItems of 500 is ten requests and takes minutes rather than seconds.
• A search returns a large total but you cannot page through all of it forever. Naver ranks results, and the tail of a broad query is padded with weak matches, so a tight query across several districts beats one loose query paged deep.
• visitor_review_score is empty for places with too few reviews to score. That is Naver declining to average three ratings, not a missing field, and it is left empty rather than filled with a zero that would drag your averages down.
• Opening hours are a single line as the business wrote it, not a structured weekly schedule. Some businesses publish none at all.
• This actor reads the place list. It does not read individual review text, menus or photo galleries from the place page.
• The site refuses some requests. The actor retries, and if every attempt is refused you get a row with blocked set and the run summary says siteRefusedUs. Treat that as "come back later", not as "no such places".

FAQ

Can I scrape Naver Map without an API key? Yes. This actor needs no key, no Naver account and no cookies from you. The official Naver Local Search API needs a registered application and returns five results per call with no review data, which is why most people end up here.

How do I get Korean restaurant data with review scores? Set businessType to restaurants and search a district in Korean, for example 강남 카페 or 홍대 맛집. Every row carries the visitor review score and both review counts.

Does Naver Map data include phone numbers? Yes, where the business publishes one. When a business only exposes Naver's forwarding number, that number is used instead, so the field is rarely empty for an active business.

Can I search Naver Map near specific coordinates? Yes. Set latitude and longitude and the search is centred on that point, so results come back ordered by distance from it.

Why does my Naver Map search return Korean text? Because the businesses are Korean and that is how Naver files them. Names, categories and addresses are stored in Korean, so translating them would mean inventing data rather than reporting it.

Related Actors

• Avito Scraper
• Google Maps Scraper
• TripAdvisor Restaurants Scraper