Minne Handmade Marketplace Scraper

Scrape handmade item listings from minne.com, Japan's largest handmade goods marketplace. Returns item name, price, creator, category, reviews, favorites, and image URLs for each product.

What it does

The scraper starts from minne's category listing pages, extracts up to 80 product IDs per page from the embedded JSON data (__NEXT_DATA__), then visits each product's detail page to collect the full record. Both listing and detail data come from minne's server-side-rendered Next.js pages — no browser or JavaScript execution is needed.

Output fields

Field	Type	Description
item_id	Integer	Minne product ID
title	String	Product name
creator_name	String	Creator or shop name
category	String	Category display name
category_slug	String	Category URL slug
price	Integer	Price in JPY
review_count	Integer	Total reviews on the creator's shop
favorite_count	Integer	Number of users who favorited this item
material	String	Product description / materials listed
image_urls	String	JSON-encoded array of product image URLs
detail_url	String	Direct link to the product page on minne.com

Input

Parameter	Type	Default	Description
maxItems	Integer	—	Required. Maximum number of items to return
categories	Array	All categories	Category slugs to scrape (e.g. ["accessories", "clothing"])

Available categories

accessories, clothing, bags_and_purses, children, interior, pets, plants, smartphone_cases_and_mobile_accessories, art, kitchen, aroma_and_candles, stationery, dolls, toys, knitting_and_needlecraft, material, kits, food

Example input

{
  "maxItems": 100,
  "categories": ["accessories", "art"]
}

To scrape all categories, omit categories or leave it empty:

{
  "maxItems": 500
}

Notes

• Language: All text (product names, descriptions, categories) is in Japanese.
• Prices are in JPY (Japanese Yen).
• Image URLs are returned as a JSON string containing an array. Parse with JSON.parse(item.image_urls).
• The actor respects rate limits and automatically backs off when minne throttles requests.
• Crawling is done via HTTP only — no proxy is required.