Kakaku.com Scraper (価格.com price watch)

This Kakaku.com scraper reads the category lists of Kakaku.com (価格.com), Japan's price-comparison site, and returns one flat row per product: no nested objects, no HTML to strip yourself. Built for Japan price monitoring and competitor tracking. Give it categories such as laptops (pc/note-pc) or vacuum cleaners (kaden/vacuum-cleaner) and you get each product's lowest price in yen, the shop that asks it, how many shops list the product, its ratings and rankings, and its release date.

Input

field	meaning
categories	One or more categories, as a path (pc/note-pc) or a list URL (https://kakaku.com/pc/note-pc/itemlist.aspx). A pasted URL keeps the filters in it (a maker, a screen size) and is paged for you
sort	bestselling (Kakaku's default), price_asc (cheapest first), price_desc, rating, reviews or newest
maxItems	Stop after this many products in total, shared across every category, in order. Kakaku shows 40 products a page
proxyConfiguration	Apify datacenter proxy by default. See Proxy below

Output

field	meaning
product_id	Kakaku's id for the row: K0001788964 for a product, J0000049506 for a group of colour variants shown as one row, all digits on old discontinued products
url, name	Product page link and the product name without the maker
maker	The maker printed before the name ("HP", "日立", "ダイソン")
variant_label	The bracketed tag next to the maker, verbatim, e.g. 【直販モデル】 (maker-direct model)
lowest_price_jpy	The lowest price any listed shop asks, in yen
lowest_price_shop	The shop that price comes from
shop_count	How many shops list the product ("全14店舗")
sales_rank	Position in Kakaku's 売れ筋 (best-seller) ranking of the category
attention_rank	Position in Kakaku's 注目 (attention) ranking
rating	Average review score out of 5, as printed
review_count	The review count printed next to the score
comment_count	The クチコミ (forum posts) count
favorite_count	How many members saved the product as a favourite
registered_date	When the product was added to Kakaku, ISO 8601
release_date_text	The release column as printed, spaces removed: 2026/5/29, 2025/8/上旬 (early August), 2026/夏 (summer), 2024/10
release_date	release_date_text as an ISO date, only when it names a full day
image_url	The product photo
color_variants	Colour families a variant-group row stands for (ブラック系, ホワイト系); empty list otherwise
spec_text	The 【スペック】 line under the row, as printed. Which specs appear depends on the category
category_path, category_name	The category the row was listed under: pc/note-pc and its Japanese name, ノートパソコン
page, position	Which page of the list the row was on, and its position on that page

The list page shows only the lowest price and a shop count. It does not show a price range or the full per-shop price list, so those are not in the output.

null is not zero

A column is null when that row of Kakaku's list does not show it. The actor never puts a 0 or an empty string in its place:

• lowest_price_jpy, lowest_price_shop and shop_count are null on a product no shop lists, where the page prints "¥―" (an arrows Tab EH on the last page of the laptop list on 2026-10-10). A 0 would read as a free product.
• rating is null on every product with no reviews: the page prints "-". The review_count beside it is 0, because the page prints "(0件)" - a printed count, not a gap.
• sales_rank and attention_rank are null where the page prints "-位", which is many products, especially on the attention ranking.
• release_date_text is null where the page prints "-". release_date is null unless the text names a full day, so "2025/8/上旬" has a text and no date.
• favorite_count is null where the favourites button shows no number.
• image_url is null where Kakaku shows its "no photo yet" placeholder.

Paging

Kakaku pages a category with pdf_pg=N and shows 40 products a page. The page prints which rows it holds ("41件～80件を表示"); the actor checks every later page starts where the last one stopped. If Kakaku ignored the request and sent an earlier page again, the actor stops and says so in the run log, and does not bill you for the repeat. A page past the end of a category is an HTTP 404; the actor reads that as "no more products", not as a failure. A product is charged once per run, even when several categories or pages list it.

Proxy

The default is Apify's datacenter proxy. On 2026-10-10 kakaku.com served full category lists (pc/note-pc, kaden/vacuum-cleaner) to both Apify datacenter and Apify residential exits. The site is behind Cloudflare, and a refusal can arrive with HTTP status 200, so the actor checks the page body, not the status. A page that is not a product list counts as a refusal: the actor switches proxy address and retries.

Billing

Pay per event: $0.00035 per product, plus $0.002 per run for actor-start. Platform usage (compute, bandwidth, proxy) is included. You pay the two event prices above and nothing else. A product listed twice in one run is charged once.

For comparison, the cheapest other Kakaku scraper on the Apify Store charges $0.0005 per item (finespun_acorn/kakaku-jp-price-ranking); the others charge $0.0026 to $0.02. This actor's $0.00035 is 30% under the cheapest.