Alibaba Product Scraper

What does Alibaba Product Scraper do?

Alibaba Product Scraper extracts wholesale product data from Alibaba.com - keyword search results and full product detail pages - anonymously, no login, no Alibaba account, no API key. Two operations live in one Actor: search for products by keyword, or look up full detail (tiered pricing, MOQ, supplier profile, specs) for one or many product IDs.

Why use this Actor?

• Sourcing/price research - search gives you ranked wholesale products for any keyword, with tiered pricing, minimum order quantity (MOQ), and supplier country/rating in one row.
• Supplier vetting - productDetails returns the supplier's company name, years on Alibaba, Gold/verified status, trade-assurance flag, and recent 6-month order volume alongside the product itself.
• Price/MOQ tracking - schedule productDetails on a cron and diff consecutive runs to catch tier-price or MOQ changes.
• Batch, one start fee - pass a list of keywords or product IDs and one run covers all of them.

How to use it

1. Pick an Operation: Product Search or Product Details.
2. Fill the matching field:
   • Search → query (e.g. bluetooth speaker) or queries for a batch.
   • Product Details → productId (e.g. 1601918386232) or a full product-detail URL, or productIds for a batch.
3. Optionally set maxItems (default 20).
4. Click Start.

Batch fields (queries/productIds) win over their singular counterparts when filled; duplicates and blanks are dropped automatically.

Input

Field	Type	Description
operation	string (required)	One of: search, productDetails.
query / queries	string / array	Search keyword(s). Used by search.
productId / productIds	string / array	A numeric product ID (e.g. 1601918386232) or a full alibaba.com/product-detail/... URL. Used by productDetails.
maxItems	integer	Caps total rows across the run. Applies to search. Default 20; first 10 rows of any run are free.
proxyConfiguration	object	Recommended: residential. Alibaba's slider-CAPTCHA bot check targets automation fingerprints more than IP reputation, but sustained volume benefits from a residential proxy.

Why does this Actor use two different techniques for its two operations?

Alibaba runs its own "Baxia"/AWSC bot-detection system, and live verification (2026-08-28/29, repeated on both a local machine and the Apify platform) found it treats the two pages this Actor uses very differently:

• A plain HTTP request (no TLS/browser fingerprint spoofing) to either page gets a 200 response whose body is entirely a <punish-component /> slider-CAPTCHA shell ("Please drag the slider to verify").
• A Chrome-TLS-impersonated plain HTTP request gets through cleanly on the product-detail page every time (10+ consecutive successes, zero CAPTCHAs) - no browser needed there at all, because the full product JSON (window.detailData) is embedded server-side in the raw HTML. A real browser, by contrast - stock headless Chromium and Camoufox (a fingerprint-patched Firefox) alike - was repeatedly blocked on this same page, both locally and on the Apify platform. So productDetails runs no browser at all: it fetches with a Chrome-impersonated TLS handshake (via apify-node-curl-impersonate) and reads window.detailData straight out of the HTML.
• The search page is the mirror image: it only server-renders a 4-item teaser and lazy-loads the rest on scroll, a plain HTTP request (impersonated or not) can't trigger that, and stock headless Chromium is hard-blocked outright with the CAPTCHA page above. Camoufox, combined with a scroll loop that triggers Alibaba's own lazy-load, gets through and returns the full result grid (48 cards confirmed after scrolling a single search page) - so search runs a real (patched) browser.

Camoufox's pass rate against the search page's challenge was not 100% in this session's live verification (one platform run returned a full 15-row result; a later run with an unrelated code change hit the CAPTCHA on every retry) - Crawlee's automatic retries (a fresh proxy session per attempt) mitigate this, but if a search run comes back empty or heavy on retries, that's this arms race, not a bug in the extraction logic; try again.

Output

One flat row per item. Search rows carry search-result fields; product-detail rows carry the full page. Every row has _operation (which mode produced it) and status (success or error) so you can split/filter a mixed export.

Search row example:

{
    "_operation": "search",
    "productId": "1601918386232",
    "title": "Premium Crystal Clear TPU PC Case for iPhone 17 18 Pro Max Strong Magnetic Charging Shockproof Transparent Phone Cover",
    "priceMin": 0.51,
    "priceMax": 0.68,
    "currency": "USD",
    "priceRangeText": "$0.51-0.68",
    "moq": 200,
    "moqUnit": "pieces",
    "soldCount": null,
    "rating": 4.7,
    "reviewCount": 104,
    "isVerifiedSupplier": true,
    "supplierName": "Guangzhou Ruite Electronic Technology Co., Ltd.",
    "supplierCountry": "CN",
    "supplierYears": 9,
    "image": "https://sc04.alicdn.com/kf/He0c23016d3854f688506c3d7090724e8a.png_300x300.png",
    "productUrl": "https://www.alibaba.com/product-detail/Premium-Crystal-Clear-TPU-PC-Case_1601918386232.html",
    "query": "phone case",
    "status": "success"
}

Product detail row example:

{
    "_operation": "productDetails",
    "productId": "1601918386232",
    "title": "Premium Crystal Clear TPU PC Case for iPhone 17 18 Pro Max Strong Magnetic Charging Shockproof Transparent Phone Cover",
    "priceMin": 0.51,
    "priceMax": 0.68,
    "currency": "USD",
    "priceRangeText": "$0.51-0.68",
    "moq": 200,
    "moqUnit": "piece",
    "rating": 4.7,
    "reviewCount": 1136,
    "isVerifiedSupplier": true,
    "supplierName": "Guangzhou Ruite Electronic Technology Co., Ltd.",
    "supplierCountry": "CN",
    "supplierYears": 9,
    "supplierProfileUrl": "https://ruitetec.en.alibaba.com/company_profile.html",
    "recentOrderAmount": "200,000+",
    "tradeAssurance": true,
    "category": ["Consumer Electronics", "Phone Case & Accessories"],
    "specifications": [{ "name": "compatible brand", "value": "For Iphone, For Samsung" }],
    "images": ["https://sc04.alicdn.com/kf/He0c23016d3854f688506c3d7090724e8a.png", "..."],
    "productUrl": "https://www.alibaba.com/product-detail/product_1601918386232.html",
    "status": "success"
}

rating/reviewCount on both operations is the supplier's store-wide rating (e.g. 4.7 / 1136 reviews), not a per-product review score - Alibaba.com is a B2B wholesale marketplace and has no per-product star-review system the way Amazon or Etsy do.

currency is read off the price text on each row, not hardcoded - Alibaba localizes the displayed currency by the visitor's (proxy) geolocation, live-verified 2026-08-29: the same product showed "$0.51-0.68" (USD) from one proxy exit and "PEN 3,458.75-3,804.63" (Peruvian Sol) from another. Expect currency to vary by run/proxy location rather than always being USD.

Pricing

This Actor uses pay-per-event pricing - one small run-start charge, then a per-row charge for each search result or product-detail lookup. See the Pricing tab for current rates. Failed lookups land as status: "error" rows and are never charged. The first 10 rows of any run are free.

Tips

• search paginates with page=N (live-verified to return distinct, non-overlapping products per page) until maxItems is satisfied - there's no reliable "last page" indicator in the DOM to stop earlier on, so set maxItems deliberately rather than leaving it unbounded for a long-tail keyword.
• For productDetails at scale, batch productIds in one run rather than one run per ID - only one run-start fee applies.

Limitations

• Only the numbers/text Alibaba shows to an anonymous, logged-out visitor are available - buyer-specific negotiated pricing is not.
• rating/reviewCount reflect the supplier's overall store rating, not this specific product.
• soldCount is only present on search rows for cards Alibaba itself shows sales history for; it is not available on productDetails.
• currency follows Alibaba's own visitor-geolocation-based localization (see Output above) - don't assume every row is USD.
• search's pass rate against Alibaba's CAPTCHA is not 100% in live testing (see the technique explanation above) - Crawlee retries with a fresh proxy session automatically, but an unlucky run can still come back with fewer rows than expected or none at all. productDetails did not show this flakiness in this session's testing (10+/10+ successes).
• Alibaba's page markup and bot-check behavior can drift over time; if a run starts returning status: "error" rows or repeated retries, re-verify against a live response.

FAQ

Does Alibaba have a public API? Alibaba.com's open API program is scoped to approved business partners, not general scraping. This Actor gets you search and product data anonymously, no approval process.

Do I need an Alibaba account or API key? No - it runs entirely against Alibaba's public, logged-out web pages.

Why does this need a real (patched) browser instead of a simple HTTP request? Only for search. productDetails uses a plain, Chrome-TLS-impersonated HTTP request - no browser - because its data is embedded server-side and a real browser was, counterintuitively, the one thing that got blocked there in live testing. search lazy-loads its full result grid on scroll and hard-blocks plain headless Chromium with a slider CAPTCHA, so a fingerprint-patched Firefox (Camoufox) is used there instead.

Disclaimer

This Actor is an independent tool, not affiliated with or endorsed by Alibaba.com or Alibaba Group. It only accesses data Alibaba.com serves publicly to anonymous visitors. Use it in compliance with Alibaba.com's Terms of Use and applicable law in your jurisdiction. Found a bug or have a feature request? Use the Issues tab on this Actor's page.