World Market scraper: home decor and furniture price, clearance and stock tracker (USA)

This World Market scraper lists World Market (the US home decor and furniture retailer) by category page or by keyword and returns one flat row per product: the current price in USD (and the high end when the product is priced as a range), the struck-through price before a markdown, the Sale and Clearance labels, in-stock status, rating and review count, sku, the product link and the photo. Run it on a schedule and diff the dataset to get a home decor and furniture price, clearance and stock watch: which dining chairs went to Clearance since last week, what a sofa line costs today, what is out of stock.

No account, no API key, no login, no browser. One request returns up to 96 products.

Input

Field	Notes
categories	World Market category pages, one per line: the page address (https://www.worldmarket.com/c/furniture/, https://www.worldmarket.com/c/bath/bath-sale/) or the category id (116715, sale-bath). A department page such as /c/furniture/ lists every product in the department (1,742 on 2026-10-21). A search page address (https://www.worldmarket.com/search?q=candle) is read as that search.
searchTerms	One or more search terms, one per line (wine glass, dining chair). Each is read separately, in the site's own order. A term with no match returns nothing and costs nothing. Terms containing < or > are rejected: World Market's firewall answers them HTTP 403.
maxItems	Products to read per category and per term (default 100, up to 10,000). 96 per page.
sort	The site's own sort: default, top-sellers, new-arrivals, rating, price-low-to-high, price-high-to-low. It decides which products a maxItems cut-off keeps.
onlyOnSale	Return only products that are marked down (on_sale) or labelled Sale or Clearance (sale_tag, clearance_tag). Other products are read, not returned and not charged, and do not count toward maxItems.
onlyInStock	Return only products with in_stock true. Others are read, not returned and not charged.
proxyConfiguration	Apify datacenter proxy by default.

At least one category or search term is required. A product that two sources both find is returned once, under the first, and charged once.

Example input:

{
  "categories": ["https://www.worldmarket.com/c/decor-and-pillows/", "sale-bath"],
  "searchTerms": ["wine glass"],
  "maxItems": 300,
  "sort": "price-low-to-high",
  "onlyOnSale": true
}

Output

Prices are numbers in US dollars; currency is USD (the en_US site).

Field	Notes
category_id, query	Where the product was found: the category id (116715, sale-bath) for a category, the search term for a search. The other is null.
position	The product's place in World Market's own order for that source (1 is first), counted over every product of the pages read. Counted before duplicates and the onlyOnSale / onlyInStock filters.
product_id	The id at the end of the product link (665798, SET131078). A string. One row is one product page, so product_id is the unique key.
sku, variant_skus	The tile's default-variant sku and every variant sku the tile lists. sku is not unique: a collection tile carries a member item's sku, so two different products can share it (295 distinct skus over 301 products of one category, all 301 links distinct). Use product_id to key rows.
name, url, image_url	As the page shows them. image_url is the tile photo with the site's own resize parameters.
categories	The site's category trail for the product, most specific first. Empty if none.
price	The price a customer pays, as the tile shows it (the sale price when the tile shows one). For a price range (a product with several variants priced differently) this is the low end. For a set (Set of 2) it is the price of the set. Null if the tile shows no price.
price_high	The high end of a price range. Null when the tile shows one price.
original_price, original_price_high	The struck-through price before a markdown (low and high end of a range). Null when the tile shows no markdown.
on_sale	True when price is below original_price (or, for a range, the high end is below original_price_high). Worked out from the prices shown.
discount_amount, percent_off	Dollars off and whole percent off. Null unless on sale and neither price is a range, never 0.
quantity	Units the listing sells together: 2 for a Set of 2. The site's own per-tile price number is per unit (half of price for a set); price is what the tile shows and what is paid.
availability, in_stock	Stock as the tile states it, lower case: in stock, out of stock or in store only; and a flag: true for in stock, false for out of stock, null for in store only (not bought online).
fulfillment	The tile's fulfilment badge: Shipping Only or Pick Up Only. Null when none.
price_label	The site's label: Sale or Clearance. Null when none.
sale_tag, clearance_tag, new_tag, best_seller_tag	The site's flags as sent. They do not always agree with the prices: a tile labelled Sale with one price and no struck-through price was seen, and Clearance tiles carry no Sale tag. on_sale and these flags are reported apart; onlyOnSale accepts either.
rating, review_count	The average star rating (0 to 5) and how many reviews it is from. rating is null for a product with no reviews (the page sends 0.00, which is not a rating of 0); review_count is 0 then.
currency	USD.

Example row (real values, 2026-10-21):

{
  "category_id": "sale-bath",
  "query": null,
  "position": 7,
  "product_id": "657923",
  "sku": "657923",
  "variant_skus": ["657923"],
  "name": "Blanca Ivory Sculpted Leaf Bath Towel",
  "url": "https://www.worldmarket.com/p/blanca-ivory-sculpted-leaf-bath-towel-657923.html",
  "image_url": "https://www.worldmarket.com/dw/image/v2/BJWT_PRD/on/demandware.static/-/Sites-wm-master-catalog/default/dw8396116a/images/large/135267_XXX_v1.jpg?sw=385&sh=385&sm=fit&sfrm=jpg&q=80",
  "categories": ["Bath Towels", "Bathroom Towels & Linens", "Bath"],
  "currency": "USD",
  "price": 9.98,
  "price_high": null,
  "original_price": 19.99,
  "original_price_high": null,
  "on_sale": true,
  "discount_amount": 10.01,
  "percent_off": 50,
  "quantity": 1,
  "availability": "out of stock",
  "in_stock": false,
  "fulfillment": null,
  "price_label": "Clearance",
  "sale_tag": false,
  "clearance_tag": true,
  "new_tag": false,
  "best_seller_tag": false,
  "rating": 5.0,
  "review_count": 1
}

Honest nulls

A field the page did not send is null, never 0 or "". variant_skus and categories are empty lists when there are none.

• No product descriptions, dimensions or per-variant prices. This is the category and search grid, one row per product page. A range shows its low and high end, not each variant.
• No stock counts. The site states in stock, out of stock or in store only; there is no quantity.
• No store-level stock. Results are the site's online view with no store selected.

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.
• $0.00069 per product row written (product-scraped), i.e. $0.69 per 1,000 products. A row that is not written is not charged: rows filtered by onlyOnSale or onlyInStock, and repeats of a product another source already returned, are free.

How it fetches

It reads the product grid that worldmarket.com's own category and search pages load (Search-UpdateGrid), an HTML fragment of product tiles whose data-* attributes and price block carry the record. Measured 2026-10-21: HTTP 200 from a plain home connection and from Apify datacenter and residential addresses (2 of 2 each), no key or cookie needed. A refusal (403, 429, 5xx), a 200 that is not a product grid, or a dropped connection is treated as the address being blocked, and the actor rotates to a fresh one and retries.

Things worth knowing, all measured:

• Paging uses start and sz, and both are honoured. The actor asks for 96 per page; start=10&sz=5 returns results 11 to 15, and sz above the match count returns everything in one body (a 301-product category, 2.6 MB). The grid carries the match count (data-total-results), and a start at or past it is HTTP 200 with no tiles, which is how a walk ends.
• A category page address is turned into its id first. The id is in the category page itself (/c/furniture/ is 116715, /c/bath/bath-sale/ is sale-bath, ids are not always numbers). The actor reads it with one extra request per category address. A category id the site does not know is answered like an empty one, so it shows as "no results", not as an error.
• The tile's own price number is not the price. It is empty on every tile priced as a range (91 of the first 200 furniture tiles) and per unit on a Set of 2. The price shown on the tile is what is read.
• sku is not a key, product_id (the id in the link) is: 1,740 of 1,740 furniture tiles had distinct links, 1,635 distinct skus.
• Sorting works and is validated. The five sorts the site offers are sent as srule; the site silently ignores an unknown one, so the actor refuses it. The price sorts follow the site's own price index, which is close to but not exactly the price shown (cheapest first, not strictly ascending in price).
• Rare unrendered tiles. 2 of 1,740 furniture tiles came back as a server-side include that failed to render, with no product in it. They are skipped and counted in the log, not returned.
• Totals drift with stock changes during a long walk (a product that moves between two pages can be read twice or missed); a product read twice is returned once.

If World Market changes its grid, the run fails with a message instead of returning empty rows.