NET-A-PORTER Scraper

Search NET-A-PORTER and get back structured product data: designer, price, the full size ladder with per-size stock, category tree, images, and the manufacturer's own style code.

Why use this actor

Most luxury scrapers hand you a name and a price. The useful part of this one is mfPartNumber, the brand's own style code, which lets you recognise the same physical garment when it appears on a different retailer's site.

If you are price-monitoring, building a resale pricing tool, or maintaining a product database across several luxury retailers, that code is the difference between "two listings that look similar" and "the same item, two prices".

What makes the style code matter

Luxury has no EAN. Groceries and electronics have a barcode that every retailer shares, so joining catalogues is trivial. Fashion does not. Each retailer invents its own SKU, and the same jacket is P01165736 at one shop and 1647597289517280 at another. There is nothing to join on.

NET-A-PORTER publishes the manufacturer's code. Example: 834141QMAC44164 on an Alexander McQueen denim blazer. That is the code McQueen itself uses (alexandermcqueen.com/.../stretch-denim-jacket-834141QMAC44164.html), and it has the structure printed on the physical hangtag: six-digit style, five-character collection and fabric, four-digit colour.

Two of the big luxury retailers expose a real manufacturer code. Farfetch does it as brandStyleId, NET-A-PORTER as mfPartNumber. NET-A-PORTER is the cheaper harvest of the two: one ordinary brand query returns 20 to 50 codes in a single response.

One honest limitation. NET-A-PORTER's search index does not tokenise these codes, so you cannot search by a style code. Searching 834141QMAC44164 returns nothing. The workflow is: search by brand and category, then match on mfPartNumber yourself. A brand query costs the same as any other and hands back a batch of codes to filter against.

How to scrape NET-A-PORTER data

1. Put one or more search terms in Search Queries. Brand plus category works best: gucci bag, mcqueen denim, loewe espadrilles.
2. Pick a Market. nap_be is Belgium in EUR; there are 19 markets and each has its own prices and stock.
3. Set Max Items and run it.

Keep queries short. NET-A-PORTER's index ANDs every token and does not stem, so each extra word can collapse the result set: mcqueen denim returns 5 hits where alexander mcqueen denim jacket returns 1, and prada shoes returns 0 where prada returns 15. The actor handles this for you (it trims stopwords, shortens long designer names, and re-asks with a shorter query when a result set comes back near-empty), but you will get better results by asking bluntly in the first place.

No login, no cookies to supply, and no proxy. The actor does the anti-bot work itself.

Input

Field	Type	Required	Description
searchQueries	Array	No	One or more search terms. Default ["gucci bag"]
searchQuery	String	No	Single-term alternative to searchQueries
storeCode	String	No	Market to read, default nap_be. See the market list below
maxItems	Integer	No	Stop after this many products across all queries. Default 50
pageSize	Integer	No	Products per search request, default 40. 50 is the ceiling
fetchProductDetails	Boolean	No	Fetch per-size availability. Default true
normalizeQueries	Boolean	No	Trim long queries to the tokens the index matches. Default true
proxyConfiguration	Object	No	Ignored on purpose, see Limitations

Markets

nap_be (Belgium), nap_nl, nap_fr, nap_de, nap_it, nap_es, nap_at, nap_ie, nap_dk, nap_se, nap_ch, nap_gb (GBP), nap_us (USD), nap_ca, nap_au, nap_hk, nap_jp, nap_kr, nap_ae.

Output

Field	Type	Example
name	String	"Denim blazer"
brand	String	"MCQUEEN"
mfPartNumber	String	"834141QMAC44164"
price	Number	1990
currency	String	"EUR"
originalPrice	Number	Only when the item is actually reduced
discount	String	"-30%", only when reduced
sizes	Array	Full ladder, one entry per SKU, with availability
inStock	Boolean	true when any size is buyable
stockBadges	Array	["LOW STOCK"], ["EXCLUSIVE"], ["Sold out"]
lowStock	Boolean	true
category	Array	["Clothing", "Jackets", "Blazers"]
categoryPath	String	"Clothing > Jackets > Blazers"
colour	String	"Blue"
imageUrl	String	Main image at 920px
imageUrls	Array	All available views
url	String	Product page on the chosen market
description	String	NET-A-PORTER's editorial copy
composition	String	"71% Cotton 29% Polyester"
sizeAndFit	String	Fit notes and model measurements
sizeSchema	String	"RTW - Italy"
productId	String	NET-A-PORTER's internal ID
partNumber	String	Variant code from the product URL
storeCode	String	"nap_be"
searchQuery	String	The query that returned this product
scrapedAt	String	ISO 8601 timestamp

Prices are converted for you

NET-A-PORTER's own API quotes money in minor units with a divisor ({"amount": 199000, "divisor": 100}). The price and originalPrice fields are already converted to major units, so 1990 means EUR 1990.00. The divisor is read from the response rather than assumed, so markets that do not use cents come out right too.

Size ladder

Each entry in sizes is one SKU:

Field	Meaning
size	Brand size label. ONE SIZE for bags and most jewellery
schemas	The same size in other scales: IT, UK, US, FR, EU, JP, AU, DE, DK, RU, KR, SML
available	Buyable right now
soldOut	Carried in this size, currently out of stock
notStockedOnline	Never carried online in this size
skuId	NET-A-PORTER's SKU id
partNumber	Per-size part number
ean	Where NET-A-PORTER publishes one, which is not always

The distinction between soldOut and notStockedOnline matters if you are tracking demand. A size that sold out was stocked and sold; a size marked notStockedOnline was never offered.

Example output

{
    "name": "Denim blazer",
    "brand": "MCQUEEN",
    "mfPartNumber": "834141QMAC44164",
    "price": 1990,
    "currency": "EUR",
    "sizes": [
        {
            "size": "40",
            "schemas": { "IT": "IT 40", "UK": "UK 8", "US": "US 4", "FR": "FR 36", "SML": "S" },
            "available": true,
            "soldOut": false,
            "notStockedOnline": false,
            "skuId": "3074457345629607675",
            "partNumber": "0400664169370",
            "ean": "813894730"
        },
        {
            "size": "42",
            "schemas": { "IT": "IT 42", "UK": "UK 10", "US": "US 6", "SML": "M" },
            "available": false,
            "soldOut": true,
            "notStockedOnline": false,
            "skuId": "3074457345629607677",
            "partNumber": "0400664169387"
        }
    ],
    "inStock": true,
    "stockBadges": ["LOW STOCK"],
    "lowStock": true,
    "category": ["Clothing", "Jackets", "Blazers"],
    "categoryPath": "Clothing > Jackets > Blazers",
    "colour": "Blue",
    "imageUrl": "https://www.net-a-porter.com/variants/images/46376663162865360/in/w920.jpg",
    "url": "https://www.net-a-porter.com/en-be/shop/product/mcqueen/clothing/blazers/denim-blazer/46376663162865360",
    "composition": "98% Cotton 2% Elastane",
    "sizeSchema": "RTW - Italy",
    "productId": "3074457345629607670",
    "partNumber": "46376663162865360",
    "storeCode": "nap_be",
    "searchQuery": "mcqueen denim",
    "scrapedAt": "2026-07-29T09:15:22.104Z"
}

Speed and cost

The run starts with a one-off browser warm-up of about 7 seconds. After that, every search page and every product detail costs roughly 0.2 seconds.

With per-size availability on, one product is one detail request. A measured run: 15 products across 3 queries in 12.6 seconds including warm-up. A 500-item run lands around 2 minutes.

Turning Fetch Per-Size Availability off removes one request per product and roughly halves the time, at the cost of the sizes ladder, description, composition and sizeAndFit.

No proxy is used, so there is no proxy cost.

Limitations

• You cannot search by style code. NET-A-PORTER's index does not tokenise mfPartNumber. Search by brand and category, then filter on the code.
• Leave the proxy off. Residential exits score worse on this site than a direct connection: measured 403 through a residential exit where direct returned 200. The setting is honoured if you turn it on, but it is an escape hatch for a burned IP range, not a default.
• The occasional run needs a second attempt. NET-A-PORTER's protection scores the exit IP, and consecutive runs from the same range do not always get the same answer. The actor retries the browser warm-up up to four times with a fresh fingerprint each time, so this normally costs you 10 to 20 seconds rather than a failed run.
• Search relevance is weak. The index ANDs tokens and does not stem, so dress finds 2,450 items but prada shoes finds none. Ask bluntly.
• Per-size availability needs the detail fetch. With fetchProductDetails: false, sizes comes back empty and inStock falls back to the sold-out badge.
• One colourway is one record. A shoe in three colours returns three records with three different style codes, which is correct: each colourway has its own manufacturer code.
• NET-A-PORTER is a full-price retailer. originalPrice and discount are usually absent because most items are not reduced.
• Data is scraped from the public website and may change without notice. Respect the website's terms of service and use responsibly.

Related scrapers

• Mytheresa Scraper — the other big European full-price luxury retailer. No manufacturer style code, so it cannot be joined the same way.
• Vinted Scraper — Europe's largest secondhand fashion marketplace. Pair it with this actor to compare retail against resale.
• Primark Scraper — the other end of the fashion price ladder.
• Rituals Scraper — luxury cosmetics and home fragrance.
• ICI PARIS XL Scraper — Belgian beauty and perfume.

Need this data on a schedule, or a custom version?

We run this scraper as a managed service for businesses: scheduled runs, deduplication, delta detection, and delivery to your inbox, Google Sheets, or API — maintenance included. We can also build a custom version with your exact fields and filters, or combine multiple luxury sources into one feed joined on the manufacturer style code.