Zid Store Scraper — Products, Prices & Stock

Paste any Zid store URL, get every public product as clean structured data. Arabic handled right, prices in the store's own currency, availability live. Free.

متجر زد | استخراج المنتجات والأسعار والمخزون

{
  "event": "product-scraped",
  "platform": "zid",
  "store": { "name": "متجر الرياض", "domain": "riyadhstore.zid.store" },
  "url": "https://riyadhstore.zid.store/products/حقيبة-جلد",
  "nameAr": "حقيبة جلد طبيعي",
  "nameEn": null,
  "price": 320.00,
  "salePrice": 279.00,
  "currency": "SAR",
  "availability": "in_stock",
  "sku": "BG-2201",
  "categories": ["حقائب", "رجالي"],
  "variants": [
    { "name": "بني / وسط", "price": 279.00, "availability": "in_stock" },
    { "name": "أسود / وسط", "price": 279.00, "availability": "out_of_stock" }
  ],
  "status": "verified"
}

Note that price and salePrice are both there. A Zid product page publishes its list price only in the storefront's own data — its structured data carries just the number a shopper pays. A scraper that reads the easy source reports a discounted product as though 279 were its normal price. This one doesn't.

---

Contents

Quick start · Why the data is trustworthy · Use cases · Input · Output · Arabic and CSV · Pricing · Limits · Errors · FAQ

---

Quick start

1. Paste one or more Zid store URLs into Store URLs — homepage, category, or product page. Mixed is fine.
2. Run it.
3. Download JSON or CSV. The CSV opens in Excel with Arabic intact.

Nothing else is required. Every other field is a refinement.

---

Why the data is trustworthy

Custom domains are detected. About 59% of Saudi Zid stores run on the merchant's own .com domain, where nothing in the URL says "zid". Detection reads the page, not the hostname, and requires several independent platform markers — so a store on yourbrand.com works exactly like one on yourbrand.zid.store, and a blog post about Zid is correctly refused with NOT_A_ZID_STORE.

Stock is never guessed. Zid storefronts can leave availability genuinely ambiguous — we measured a sold-out product whose stock flag was empty. This Actor reads both the storefront's own state and its structured data. When they agree, you get the answer. When they disagree, you get unknown and a warning — never a confident wrong answer. A stock field you cannot believe is worse than one that admits doubt.

Unreadable means null. No field is ever filled with a guess. Anything uncertain raises a warning and degrades status from verified, so you can filter on it.

Every row is self-contained, so an agent can act on one row without fetching anything else.

The scorecard

Measured on 38 live Zid stores, 152 products — not a demo store, and not a list picked for being easy to parse:

Field	Success
Product name	100% (152/152)
Price	100% (152/152)
Currency	100% (152/152)
Availability	100% (152/152)
SKU	100% (152/152)
Images	100% (152/152)
Variants	100% (30/30 products that have options)
Categories	96.7% (147/152)

28 of the 38 stores run on their own domain, where nothing in the URL says "zid" — and every one was detected. No store was misidentified in either direction. Their combined catalogs come to over 56,000 products.

One store of the 39 listed could not be read at all, and it is kept in the list rather than dropped because it marks the real edge of what this Actor can do. It answers sitemap_products.xml with an HTTP 400 and a 220 KB HTML error page, so the category-crawl fallback takes over — and finds nothing, because that storefront builds its product grid in JavaScript. Its category pages contain one link, /products/, and no individual products.

So the fallback rescues a store whose sitemap is missing but whose pages are ordinary HTML. It cannot rescue one that renders its catalogue client-side; reading that store means running a browser, which this Actor deliberately does not do. The run says so rather than returning an empty catalogue as though the shop were bare: discoveryMethod reads category-crawl, the row carries a SITEMAP_UNAVAILABLE warning, and no products are claimed.

The five products missing categories are counted as misses even though a merchant filing a product in no category is perfectly normal — we did not verify which it was, and a scorecard should round against itself.

On currency. Every product row carries its own currency, and that is the number this Actor stands behind. What it deliberately does not promise is that the currency will be SAR. The sibling Salla Actor measured a storefront returning 290 SAR read from Saudi Arabia and 105 USD read from a US data centre — a different currency and a different number for the same product, because the storefront localises on the caller's address, and no query parameter or header would override it. Whether Zid storefronts do the same has not been measured here, and where a platform runs from is not something a scraper controls, so the listing claims what is true either way. If you need prices as a Saudi shopper sees them, point proxyConfiguration at a residential proxy in SA (apifyProxyCountry) and the store will localise to that instead.

Every figure above is a count from that run, not an estimate, and the measuring tool is kept in the source tree beside the Actor rather than written up after the fact. Ask on the Actor's issues page if you want the store list and the probe script to reproduce it.

---

Use cases

You are	You get
A merchant on Zid	A backup or export of your own catalog — for migration, an audit, or a feed
A competing merchant or agency	Competitor prices and assortment across KSA
A market researcher	Structured MENA e-commerce data
Moving between platforms	Your catalog out of one storefront and into another

Price-drop watching: feed the url column into the Website Change Monitor and get told when a competitor moves a price.

---

Input

Field	Type	Default	Meaning
startUrls	list of URLs	—	Zid store home, category, or product URLs
maxItems	integer	200	Stop after this many products per run
includeOutOfStock	boolean	true	Keep sold-out products
includeVariants	boolean	true	Expand publicly visible options
language	auto / ar / en	auto	Storefront language to expect — checked, not switched (see below)
respectRobots	boolean	true	Skip what robots.txt disallows
proxyConfiguration	object	—	For stores behind aggressive CDNs

On language. This Actor reads whatever the storefront serves; it does not translate and it does not switch the shop into another language. Setting ar or en states what you expect, and the store row carries a LANGUAGE_UNAVAILABLE warning when the shop does not publish it — so a catalogue that came back in the other language says so rather than looking like the one you asked for. Stores that declare no language at all are left alone: silence is not a denial.

---

Output

Two record types share one dataset, told apart by event. Everything above is real output from a live run against four Zid stores — no sample data was written by hand, and the one product page that timed out is shown exactly as the dataset carries it.

product-scraped

One row per product. Fields: platform, store, url, nameAr, nameEn, price, salePrice, currency, availability, sku, brand, categories, images, variants, scrapedAt, status.

availability is in_stock, out_of_stock, or unknown. images are URLs — this Actor never downloads images.

store-analyzed

One row per store per run, carrying the summary:

{
  "event": "store-analyzed",
  "platform": "zid",
  "store": { "name": "متجر الرياض", "domain": "riyadhstore.zid.store" },
  "productsDiscovered": 187,
  "productsScraped": 187,
  "stoppedBy": "completed",
  "discoveryMethod": "sitemap",
  "warnings": []
}

stoppedBy is completed, maxItems, timeLimit, or error. When productsDiscovered exceeds productsScraped, the catalog is bigger than your maxItems — raise it and run again. discoveryMethod is sitemap, category-crawl, or direct-url.

platform and the sister Actor

Every row carries platform. The Salla Store Scraper uses identical field names and the same values inside them, so a Zid dataset and a Salla dataset concatenate into one sheet with no renaming, and platform still tells you where each row came from. The values matter as much as the names: a filter on stoppedBy = maxItems has to catch both Actors' rows or it quietly returns half an answer.

Concretely: every column a Zid row publishes, a Salla row publishes under the same name. The reverse is not quite true, and deliberately so — a Salla row carries five extra columns for measurements only that platform needs (hasOptions, finalUrl, productsWithOptions, pricesCrossChecked, pricesDisagreed). Merging the two leaves those blank on Zid rows rather than requiring you to rename anything. Inventing Zid equivalents would mean publishing columns that are always zero, which reads as a measurement rather than an absence.

That is what makes the migration case work: export a catalogue from one platform, export from the other, stack them, and sort by platform.

---

Arabic and CSV

Product names, categories, and variant options come back in Arabic exactly as the merchant wrote them. CSV exports carry a UTF-8 BOM so Excel opens them with Arabic intact rather than as mojibake.

An Arabic product name containing a Latin model number — سماعة رأس هايبر أكس 727A9AA — stays in nameAr whole. It is not split. Stores publish one name, so the other field stays null rather than holding a transliteration nobody wrote.

---

Pricing

Free. Only Apify's platform usage applies. Two events are defined — product-scraped (one per product) and store-analyzed (one per store per run) — and neither is charged today.

They exist so that adding a price later would be a price change and not a redesign. Nothing is planned.

---

Limits

Limit	Value
Store URLs per run	50
Products per run (maxItems)	200 default, 10,000 maximum
Products discovered per store	50,000
Images per product	25
Categories per product	25
Variants per product	200
Per-page timeout	15 seconds
Per-store timeout	600 seconds
Stores in parallel	2

The discovery ceiling is far above maxItems on purpose: we want to report that a catalog is bigger than your run rather than let a truncated list look like a small shop.

---

Errors

Code	Meaning
NOT_A_ZID_STORE	Reachable, but not a Zid storefront
ROBOTS_DISALLOWED	robots.txt asks us not to crawl this; respected by default
STORE_UNREACHABLE	The store could not be reached — retryable
STORE_TIMEOUT	The store did not respond in time — retryable
PRODUCT_PARSE_FAILED	One product was unreadable; the run continues
INVALID_URL / INVALID_INPUT	The input could not be used
BLOCKED_DESTINATION / DNS_FAILURE	The address is not a public website
BUDGET_EXHAUSTED	The run's item budget ran out
STORAGE_FAILURE / INTERNAL_ERROR	Something on our side

A single unreadable product never ends the run. It comes back as a row with status: failed, so you can see which URL did not work instead of finding a silent hole in your data.

---

FAQ

Does it work on a store with its own domain? Yes — that is the majority of Zid stores, and detection is built for it.

Does it need my store login? No. Public catalog pages only. This Actor never touches logins, carts, checkout, or coupons.

Does it get reviews or customer data? No. Product data only. Zid's own robots.txt asks crawlers to leave review pages alone, and we do.

Does it download the images? No — you get URLs.

Does it scrape Salla stores? No. That is the separate Salla Store Scraper, with identical field names.

Can it find every Zid store for me? No. It takes the store URLs you give it.

---

Support

Open an issue on the Actor's Apify page. Include the run ID and the store URL — the store-analyzed row tells us most of what we need.

Not affiliated with, endorsed by, or sponsored by Zid.