⭐ Coupang Eats Reviews Scraper

Extract customer reviews — ratings, review text, photos, ordered menu items, reorder signals, and owner replies — from Coupang Eats (쿠팡이츠), South Korea's #2 food delivery platform.

This is the companion actor to the Coupang Eats Crawler (restaurants + menus). That one answers "which restaurants exist and what do they sell?" — this one answers "what do customers actually think?"

---

🌏 What is Coupang Eats?

Coupang Eats is the food-delivery arm of Coupang, Korea's largest e-commerce company (the "Amazon of Korea", NYSE-listed CPNG). Together with Baemin it controls ~88% of the Korean delivery market, and it's the side still gaining share — every Coupang Wow membership (~15M subscribers) bundles free Eats delivery.

Reviews on Coupang Eats are order-verified: a customer can only review a store after actually ordering from it. That makes this review corpus far cleaner than open platforms like Google Maps — no drive-by one-stars, no competitor sabotage, and every review is linked to the actual menu items ordered.

---

🚀 What does this actor do?

Give it a store (by ID or by name) and it returns that store's reviews, newest-first by default:

• Store IDs (storeIds) — the storeId from the share URL. Fastest, no address needed.
• Store names (storeNames) — Korean or English. Matched to a store near your chosen address (results depend on the delivery location; chains have one store per district, so set the address near the branch you mean).
• Choose sort order (newest / most helpful / highest / lowest rating), cap reviews per store, optionally keep only photo reviews.

No account or extra setup required — just enter a store and run.

---

🍗 Why review data matters (buyer personas)

Buyer	What they use the data for	Refresh cadence
Restaurant brands & franchises	"What do customers complain about at our Gangnam branch vs Hongdae? Which menu items drive 1-star reviews?"	Daily / weekly
Reputation management & CX tools	Track owner reply rates and response times; alert franchises on unanswered negative reviews.	Daily
F&B market research	Sentiment by cuisine and district; which dishes are praised, which complaints recur (cold food, missing items, portion size).	Weekly
Menu intelligence	Reviews link to the exact menu items ordered — mine which dishes get reordered ("N번째 재주문") and co-ordered.	Weekly
Hedge fund / alt-data desks	Review velocity as a demand proxy per store / district / brand — a leading indicator on CPNG and Woowa/DH.	Weekly
Restaurant SaaS	Lead-scoring: stores with high volume but low ratings or no owner replies are the best prospects for CX tooling.	Monthly

---

🗾️ Example input

By store ID (fastest, no address needed):

{
  "storeIds": ["741250"],
  "maxReviewsPerStore": 100
}

By store name (matched near an address):

{
  "storeNames": ["무궁화반점 강남점", "맥도날드"],
  "addresses": [
    { "latitude": 37.4979, "longitude": 127.0276, "label": "Gangnam, Seoul" }
  ],
  "maxReviewsPerStore": 200
}

Worst reviews first, photos only:

{
  "storeIds": ["741250"],
  "sort": "RATING_ASC",
  "onlyWithPhotos": true,
  "maxReviewsPerStore": 50
}

---

📦 What data do you get?

One row per review:

{
  "review_id": "273228948",
  "store_id": "741250",
  "store_name": "무궁화반점 강남점",
  "rating": 5,
  "text": "근래 먹은 짬뽕 중 가장 맛있었어요~~",
  "writer": "주*연",
  "written_at_text": "오늘",
  "written_date_approx": "2026-06-10",
  "images": ["https://t4c.coupangcdn.com/thumbnails/remote/1024x1024/image/eats_review_api/....jpg"],
  "image_count": 1,
  "ordered_menu_items": ["소고기 직화 짬뽕", "500원의 행복:튀김고기만두 2개"],
  "thumb_up_count": 0,
  "reorder_count": 2,
  "merchant_reply_text": "안녕하세요, 무궁화반점 강남점입니다 😊 ...",
  "merchant_reply_written_at_text": "오늘",
  "writer_review_count": 2,
  "writer_rating_avg": 5.0,
  "is_owner_review": false,
  "store_rating_avg": 4.8,
  "store_review_count": 6851,
  "store_rating_distribution": { "5": 87, "4": 8, "3": 3, "2": 1, "1": 1 },
  "sort": "LATEST_DESC",
  "review_rank": 1,
  "input_store_name": null,
  "store_url": "https://web.coupangeats.com/share?storeId=741250",
  "scraped_at": "2026-06-10T08:15:00.000Z"
}

Field reference

Field	Type	Notes
review_id	string	Unique Coupang Eats review ID
store_id, store_name	string	The reviewed store
rating	number	Star rating (1–5)
text	string	Review body (Korean)
writer	string	Reviewer name, masked by the platform (e.g. 주*연)
written_at_text	string	Relative date as shown in the app (오늘, 3일 전, 1주 전, 2개월 전)
written_date_approx	string	The relative date converted to an approximate YYYY-MM-DD
images, image_count	array / number	Review photo URLs
ordered_menu_items	array	Names of the menu items this customer actually ordered
thumb_up_count	number	"Helpful" votes from other customers
reorder_count	number	Set when the platform shows "N번째 재주문" (Nth reorder) — a strong loyalty signal
merchant_reply_text	string	Owner's public reply (null when the owner didn't reply)
merchant_reply_written_at_text	string	Relative date of the owner reply
writer_review_count, writer_rating_avg	number	This reviewer's lifetime review count and average rating on the platform
is_owner_review	bool	Platform flag (rare)
store_rating_avg, store_review_count	number	Store-level summary at scrape time
store_rating_distribution	object	Percent of reviews per star level, e.g. {"5": 87, "4": 8, ...}
sort	string	The sort order this run used
review_rank	number	Position of the review under that sort
input_store_name	string	The name you searched for (when the store was resolved from storeNames)
store_url	string	Public share URL
scraped_at	string	ISO timestamp of the run

Note on dates: Coupang Eats only exposes relative dates ("3일 전"). written_date_approx converts them to calendar dates; precision degrades with age (weeks → ±3 days, months → ±2 weeks).

---

🦖 How to use

1. Enter storeIds (find the ID in any Coupang Eats share link: https://web.coupangeats.com/share?storeId=741250) — or storeNames + an address.
2. Optionally set maxReviewsPerStore (default 100), sort, and onlyWithPhotos.
3. Run the actor.
4. Download from the Dataset tab or via the API.

A run collecting 100 reviews from one store completes in a few seconds.

Tip: to scrape reviews for many stores at once, first run the Coupang Eats Crawler with a search/category, export the store_id column, and paste it into this actor's storeIds.

---

🔍 Key features

• Order-verified reviews — every review is tied to a real order, with the exact menu items listed.
• Owner replies included — measure reply rate and tone per store.
• Loyalty signals — reorder counts and reviewer history (lifetime review count + average rating).
• All 4 app sort orders — newest, most helpful, highest, lowest.
• Photo filter — collect only reviews with images.
• Two input modes — direct store IDs (no address) or store names matched near an address.
• Automatic retries — transient errors retried with exponential backoff.

---

📋 Input parameters

Parameter	Type	Default	Description
storeIds	array of strings	[]	Numeric store IDs. No address required.
storeNames	array of strings	[]	Store names to resolve via search — requires addresses.
addresses	array of {latitude, longitude, label}	Gangnam	Delivery point used for name resolution (results depend on location).
maxReviewsPerStore	integer	100	Cap per store.
sort	string	LATEST_DESC	LATEST_DESC (newest), LIKE_DESC (most helpful), RATING_DESC, RATING_ASC.
onlyWithPhotos	boolean	false	Only reviews with photos.
proxyConfiguration	object	none	Optional. Not required for normal runs.

Common address coordinates (for name lookup)

District	Latitude	Longitude
Seoul — Gangnam Station	37.4979	127.0276
Seoul — Hongdae	37.5563	126.9236
Seoul — Itaewon	37.5347	126.9947
Busan — Seomyeon	35.1796	129.0756
Jeju City	33.4996	126.5312

---

📩 Feedback

Found a bug or have ideas? Open an issue on the actor's Apify page — happy to improve it.