Naver Place Reviews + Blog — Korean Review Scraper

Extract visitor reviews and blog reviews from Naver Place — Korea's #1 local business platform. Get full review text (Korean), star ratings, reviewer profiles, visit details, voted keywords, photos, owner replies, linked Naver Blog posts, and aggregated review statistics. No login, no API key, no paid proxy required.

What this actor does

• Visitor reviews — full review body, 1–5 star rating, reviewer nickname + profile URL + total review count, review/visit dates, visit count, voted keywords (음식이 맛있어요, 분위기가 좋아요…), photos, menu item mentioned, owner reply
• Blog reviews — Naver Blog posts linked to the place: title, contents, post URL, blogger name + home, thumbnail, rank
• Review statistics — average rating, total review count, star distribution (5★→1★), review themes, voted-keyword analysis
• Two input modes — direct Place URLs/IDs or keyword search to discover places first
• Parallel place scraping — configurable concurrency with per-place retries
• Empty fields are omitted — every record only contains fields Naver actually returned

How it works

Naver Place's review page (pcmap.place.naver.com/place/<id>/review/visitor) fully server-side renders ~20 visitor reviews, linked blog reviews, and review statistics into its window.__APOLLO_STATE__ blob. This actor parses that state directly with curl_cffi Chrome TLS impersonation — no browser, no login, no captcha, no proxy needed for moderate traffic.

Note on pagination: Naver's cursor-pagination GraphQL API (pcmap-api.place.naver.com/graphql) is protected by an interactive captcha against automated requests, so this actor collects the reviews Naver embeds in the SSR page (≈20 per place). This matches the reliability-first design of this repo's other Naver actors (see NaverMapScraper).

Input

Field	Type	Default	Description
placeUrls	array	—	Naver Place URLs or bare Place IDs. Accepts map.naver.com/p/entry/place/<id>, pcmap.place.naver.com/place/<id>/..., m.place.naver.com/..., or just the numeric ID.
searchKeywords	array	—	Keywords to find places first (e.g. 강남 맛집, Seoul BBQ).
maxPlacesPerKeyword	int	3	Max places scraped per keyword (1–50).
maxReviews	int	20	Max visitor reviews collected per place (SSR embeds ~20).
includeBlogReviews	bool	true	Also emit linked Naver Blog reviews.
includeReviewPhotos	bool	false	Include photo URLs on each visitor review.
includeReviewStats	bool	true	Emit one stats record per place.
maxConcurrency	int	3	Places scraped in parallel (1–10).
maxItems	int	100	Hard cap on total records (1–5000).
proxyConfiguration	object	AUTO	Optional proxy.

Example: fetch reviews for specific places

{
  "placeUrls": [
    "https://map.naver.com/p/entry/place/38239053",
    "1000750444"
  ],
  "includeReviewPhotos": true,
  "maxItems": 50
}

Example: discover places by keyword, then fetch reviews

{
  "searchKeywords": ["강남 맛집", "홍대 카페"],
  "maxPlacesPerKeyword": 3,
  "includeBlogReviews": true,
  "includeReviewStats": true,
  "maxItems": 100
}

Output

Each record has a reviewType field: visitor, blog, or stats.

Visitor review record

Field	Description
reviewType: "visitor"	Record kind
placeId, placeName, placeCategory, placeAddress, placeRoadAddress	Place identity
placeRating, placeTotalReviews	Place-level rating & count
placeUrl	Canonical place URL
searchKeyword	Keyword that found this place (search mode only)
reviewId	Naver review ID
reviewText	Full review body (Korean)
reviewRating	1–5 stars
reviewerName, reviewerProfileUrl, reviewerReviewCount	Reviewer identity
reviewDate, visitDate	Review creation & visit dates (YYYY-MM-DD)
visitCount	"n번째 방문" count
votedKeywords	e.g. 음식이 맛있어요, 분위기가 좋아요
reviewPhotos	Photo URLs (when includeReviewPhotos)
reviewReplyText, reviewReplyDate	Business owner's reply
menuItem	Menu item mentioned in the review
originType	e.g. 영수증 (receipt-verified)
language, translatedText	Review language + auto-translation when available
scrapedAt	ISO 8601 timestamp

Blog review record

Field	Description
reviewType: "blog"	Record kind
placeId…placeUrl	Place identity (same as above)
blogTitle	Blog post title
blogContents	Blog post excerpt/contents
blogUrl	Link to the full blog post
bloggerName, bloggerHome	Blogger identity
blogThumbnailUrl	Post thumbnail
blogRank	Position in Naver's blog review list
scrapedAt	ISO 8601 timestamp

Stats record (one per place)

Field	Description
reviewType: "stats"	Record kind
averageRating	Average visitor review score
totalReviewCount	Total visitor reviews
starDistribution	{5: n, 4: n, 3: n, 2: n, 1: n}
reviewThemes	[{label, count}] — theme analysis
votedKeywordStats	[{keyword, count}] — most-voted keywords

Use Cases

• Korean market entry research — what locals say about competitors
• Restaurant/cafe competitive intelligence — ratings, keywords, sentiment
• Customer sentiment analysis — NLP-ready structured Korean review data
• Brand reputation monitoring in Korea
• Academic research on Korean consumer behavior

FAQs

Is a Naver login or API key required? No. The review page is public and server-side rendered.

How many reviews can I get per place? Naver embeds ~20 visitor reviews per SSR page. Naver's deeper pagination API is captcha-protected against automated requests, so this actor intentionally collects the SSR-embedded set (matching the reliability-first design of this repo's Naver actors).

Why is blogDate missing? Naver does not expose a publish date for blog reviews in the review SSR data, and the blog post ID in the URL is not a timestamp — emitting a wrong date would be worse than none.

Can I get reviews for places I don't know yet? Yes — use searchKeywords to discover places first (e.g. 강남 맛집), then reviews are scraped automatically.

What are the Korean fields? votedKeywords are Naver's official review-vote tags (음식이 맛있어요 = "food was delicious", 분위기가 좋아요 = "nice atmosphere"); originType: 영수증 marks receipt-verified reviews.

Does this actor work with the existing NaverMapScraper? Yes — use NaverMapScraper to build a place list (search/detail enrichment), then feed the place IDs into this actor's placeUrls for full review extraction.

Data Source

All data comes from the public Naver Place review pages (pcmap.place.naver.com) — a free, publicly accessible source requiring no login or API key.