Hwahae K-beauty Rankings Scraper (Korea)

Read the bestseller rankings of Hwahae (화해), Korea's largest cosmetics review app, and get one flat row per ranked product: the rank, how far it moved since the last ranking, whether it is new to the list, the product and brand, the shop price in KRW, the discount, the rating and the review count. Run it on a schedule and keep the dataset to get a daily K-beauty trend history: what is climbing in skincare, what a 20s shopper buys, which brand just entered the top 50.

No account, no API key, no login, no browser.

What you can read

Hwahae has four product ranking themes, each with a tree of categories:

Theme	Rows	What it ranks
trending (급상승)	50	Products rising fastest, by category. Rebuilt daily at 05:00 Korea time.
category (카테고리별)	100	Best products per category (skincare, cleansing, mask, sun care, makeup, body, hair, perfume, men ...).
skin (피부별)	100	Best products per skin type (dry, oily, normal, combination, sensitive, acne, atopic), then by category.
age (연령대별)	100	Best products per age band (10s, 20s, 30s, 40s+), then by category.

Hwahae also has a brand ranking; it ranks brands, not products, and is not supported.

Input

Field	Notes
rankings	One ranking per line. Default: trending. See below.
maxItems	Rows per ranking, best rank first (default 100, which is the whole ranking; trending holds 50).
proxyConfiguration	Apify datacenter proxy by default.

Naming a ranking

Write a theme, or a path down the theme's category tree, steps joined with >:

trending
trending > 스킨케어
skin > 민감성 > 스킨케어
age > 20대
category > 남성

• A group means its overall ranking (trending > 스킨케어 is the ranking of all of skincare). Names under the group go as deep as Hwahae's tree goes (trending > 스킨케어 > 스킨/토너 > 수분).
• The tree is Korean only. The top groups, skin types and age bands also take English names: trending > skincare, skin > sensitive, age > 20s, category > men. Case and spaces are ignored.
• skin and age on their own read Hwahae's default for that theme (what the site opens on: combination skin, 30s). The ranking_name column says which.
• A name that is not in the tree fails that ranking (no charge) and the log lists what the valid names are at that step.
• Instead of a path you can paste the ranking's page URL from hwahae.co.kr (https://www.hwahae.co.kr/rankings?english_name=skin&theme_id=687), or write the theme and its numeric id: skin > 687.
• A ranking asked for twice (by name and by id, say) is read once.

Output

One row per position in a ranking. The same product appears once in each ranking it is in, so the key of a row is ranking_id + rank.

Field	Notes
theme, theme_name, ranking_id, ranking_name, ranking_url	Which ranking the row belongs to. ranking_name is the path in Hwahae's tree (Korean), e.g. 스킨케어 > 전체.
ranking_updated_at, ranking_size	When Hwahae last rebuilt the ranking (as sent, Korea time, no offset) and how many rows it has.
rank, rank_delta, is_rank_new	Position from 1. rank_delta is places moved since the previous ranking: positive = up, negative = down, 0 = same. is_rank_new marks a product that was not in the previous ranking.
product_id, product_name, product_url, image_url	Hwahae's catalogue entry.
brand_id, brand_name	As Hwahae names the brand (Korean).
price, currency, discount_rate, capacity	The shop price in KRW of Hwahae's own offer for the product, the percent off as Hwahae states it (the reference price it is measured against is not in the data) and the offer's size.
offer_id, offer_name, offer_url, offer_image_url	Hwahae's shop item. Its title often carries promo text such as [only화해] ... 더블 기획.
catalog_price, package_info	The price and pack size on the catalogue entry.
rating, review_count	Hwahae review score out of 5 and review count.

Example row

Trending, rank 1 on 2026-10-14:

{
  "theme": "trending",
  "theme_name": "급상승",
  "ranking_id": 5102,
  "ranking_name": "카테고리 전체",
  "ranking_url": "https://www.hwahae.co.kr/rankings?english_name=trending&theme_id=5102",
  "ranking_updated_at": "2026-10-14T05:00:00",
  "ranking_size": 50,
  "rank": 1,
  "rank_delta": 0,
  "is_rank_new": false,
  "product_id": 1990260,
  "product_name": "파워10 감초줄렌 젤리패드",
  "product_url": "https://www.hwahae.co.kr/products/1990260",
  "catalog_price": 25000,
  "package_info": "70EA(120ml)",
  "brand_id": 56,
  "brand_name": "잇츠스킨",
  "offer_id": 43022,
  "offer_name": "파워10 감초줄렌 젤리패드(대용량) 130매",
  "offer_url": "https://www.hwahae.co.kr/goods/43022",
  "price": 9500,
  "currency": "KRW",
  "discount_rate": 76,
  "capacity": "130EA/215ml",
  "rating": 4.65,
  "review_count": 7301
}

Honest nulls

A field Hwahae did not send is null, never 0 or "".

• Many products have no shop offer (223 of 900 rows sampled across 14 rankings on 2026-10-15; the share varies by ranking and is higher in narrow categories). Hwahae ranks it from its catalogue and reviews but does not sell it. On those rows price, discount_rate, capacity and every offer_* field are null; catalog_price is still there.
• price and catalog_price are different things. The shop sells bundles, refills and double packs, so the shop price can be above or below the catalogue price of the single item (higher on 241 of 677 offers sampled). Do not compute a discount from the pair; use discount_rate.
• catalog_price is null for the rare product whose catalogue entry has no price (3 of 900 sampled). capacity is null when Hwahae sends an empty one.
• rating is null for a rating of 0, which would mean "not rated".
• rank_delta is whatever Hwahae sent. Every NEW product sampled carried a rank_delta of 0, so read is_rank_new and rank_delta together.

Pricing

Pay per event:

• $0.002 per run (actor-start), charged only after your input is validated, so a run that fails on bad input costs nothing.
• $0.0035 per ranking row written (ranking-row-scraped), i.e. $3.50 per 1,000 rows. The default run (trending, 50 rows) costs about $0.18. A row that is not written is not charged.

How it fetches

The ranking page hwahae.co.kr/rankings?english_name=<theme>&theme_id=<id> is server rendered: it carries the theme's whole category tree and the first 20 rows of the ranking. The tree is what lets you name a ranking instead of knowing its id. Rows 21 and on are the ones the page itself loads as you scroll, from Hwahae's own JSON gateway (gateway.hwahae.co.kr), 20 per request. Both answer without a login. A refusal (403, 429, 202, 5xx), a 200 that is not the page or the JSON, is treated as the address being blocked: the actor rotates to a fresh one and retries.

Things worth knowing, all measured on 2026-10-15:

• Only the most specific category is a ranking. A group such as 스킨케어 has no ranking of its own; its child 전체 ("all") is the ranking of the whole group. Naming the group picks that child for you. Hwahae's own site answers an id that is a group with an error page (HTTP 500), and so does this actor (as a failed ranking, no charge).
• A wrong id is an error, never a refusal. An id that does not exist is HTTP 404 and a made-up theme is HTTP 500, both on Hwahae's own error page; the actor does not retry them. An id from another theme is served by Hwahae with that other theme's rows under this theme's tree; the actor checks the id is in the theme's tree and fails the ranking rather than label the rows wrong.
• Update rhythm differs by theme. On 2026-10-15 trending had been rebuilt on 2026-10-14 at 05:00, and the category, skin and age rankings on 2026-10-08. Read ranking_updated_at before treating two runs as two days of movement.
• Ranking length. Trending holds 50 rows; the others 100. maxItems above a ranking's length reads the whole ranking.
• Search is not covered. Hwahae's search pages are rendered in the browser and are not part of this actor.
• Brand rankings are not covered. They rank brands, not products.