Alibaba Scraper: Products, Prices, MOQ, Suppliers

Alibaba Scraper: Products, Prices, MOQ, Suppliers

Alibaba scraper for product offers with the real price at the minimum order: search Alibaba.com by keyword, search link, category or product link and get one row per offer with the USD price per unit at the minimum order quantity (MOQ), the quantity price ladder, MOQ and unit, a per piece flag, and the supplier's facts (country, years on Alibaba, Verified, Verified Pro, Trade Assurance, business type, response rate, store rating). Every search adds one price summary row with the median and quartiles of that price. No login, no cookies, no browser.

It is built for Amazon private label and FBA sellers sizing a product idea, sourcing agents building supplier shortlists, product and market researchers comparing factory prices across keywords, and apps and AI agents that need a typical wholesale price for a product in one call.

What it does

• The price at the minimum order, not the bulk price: unitPriceAtMoq is the ladder tier that holds the MOQ. The low end of an offer's range (priceMin) is the price at the largest order and is kept apart.
• The price at your quantity: give orderQuantity and every row adds unitPriceAtQuantity from the offer's own ladder.
• A clean unit on every row: priceUnit (piece, set, pair, meter, kilogram and 17 more), isPerPiece and unitClass, so a price per meter never mixes with a price per piece.
• A price summary per search: a headline median and quartiles of the price at MOQ (medianUsd: offers whose title holds every keyword word, else the first offers Alibaba ranks best inside your price range), and beside it the median, quartiles, minimum and maximum per piece and per sellable item, over the offers delivered.
• Alibaba's own filters, checked again on every row: price, MOQ, supplier country, Verified, Verified Pro, Ready to Ship, certifications, store rating, plus filters applied on the row: Trade Assurance, business type (manufacturer and four more), years on Alibaba, lead time, ads, title match and ids to leave out. A row that fails a filter is left out, free.
• Pasted links: Alibaba search, catalog and product links, read with their own filters.
• Product details on request: attributes, variants and their prices, customisation, sample price, lead times, packaging, places of dispatch and certifications with validity.
• A search that matches nothing is free: Alibaba answers every word, nonsense included, with 48 offers. A search whose first page holds no offer whose title carries the keyword delivers one free no_match row and no billed offer.

Quick start

1. Open the Actor and type one or more keywords as you would on Alibaba.com, for example yoga mat or

   steam cleaner handheld
   .
2. Set maxResultsPerKeyword (20 by default, up to 960) and add filters if you want: a price range, a largest MOQ, supplier countries, Verified suppliers only.
3. Turn on orderQuantity if you plan a given order size, and includeDetails if you need the product page facts.
4. Run it, then download the dataset as JSON, CSV or Excel, or read it through the API.

The prefilled input searches yoga mat and delivers 20 offers and one price summary row in a few seconds.

What you get

One row per offer (type: "offer"), and per search one price summary row (type: "keywordSummary") and one coverage row (type: "search": pages read, offers seen and delivered, repeats, rows left out by reason, why the walk ended). A product link gives one type: "product" row. Status rows (type: "status") name anything not delivered (an input problem, a product Alibaba no longer shows, a search with no match, a stop), and the run summary (type: "summary") comes last. Only offer and product rows are billed as rows; the summary, coverage and status rows are free.

Fill rates are measured over 10,560 offers of 11 keywords read on pages 1 to 20 (2026-11-04); the unitPriceAtMoq figure is from 200 offers of 10 keywords checked against their product pages:

Field	What it holds	Example (platform run, yoga mat)	Filled
offerId, title, url, imageUrl	the offer	1601799200051	100%
unitPriceAtMoq	USD price per unit at the minimum order	5.12	76% (the rest price by variant or state no single price)
priceBasis	quantity_ladder, single_price, price_range or not_stated	quantity_ladder	100%
priceTiers	the quantity ladder: minQuantity, maxQuantity, priceUsd, promotionPriceUsd	20 to 499: 5.12, 500+: 4.72	67.9% (offers with 2 or more tiers; a one tier offer shows its single price in unitPriceAtMoq)
priceMin, priceMax	the offer's USD range; priceMin is the price at the largest order	4.72, 5.12	100%
priceText	Alibaba's own price text	US $4.72-$5.12	100% (in US $ 99.8%)
moq, moqUnit, moqText	the minimum order as a number, its unit and Alibaba's text	20, piece, 20 pieces	100%
priceUnit, isPerPiece, unitClass	the normalised unit; true for piece or unit; item, length, area, weight, volume, pack or other	piece, true, item	100% (per piece on 66.0%)
leadTimeDays	the lead time on the search card	7	100%
leafCategoryId, leafCategoryName	Alibaba's category	Yoga Mats	100%
supplier.name, supplier.id, supplier.country, supplier.years	the supplier	CN, 1	100%
supplier.verified, supplier.verifiedPro, supplier.goldSupplier	Alibaba's badges	false, false, true	100%
supplier.tradeAssurance	takes Trade Assurance orders	true	true on 98.7%
supplier.businessTypes	manufacturer, trading_company, distributor, buying_office, other	["trading_company"]	98.6%
supplier.responseRate	response rate in percent	84.4	99.3%
supplier.rating, supplier.reviewCount	store rating out of 5 and review count	4.8, 2	74.6%
unitsSold, views	sold and view counts where Alibaba shows them	null	23.3%, 4.7%
isAd	a paid placement	false	true on 14.4%
titleMatchesKeyword, titleMatchShare	the title holds at least half the keyword's words	true, 1	100% (the share is empty on category browses)

Also on the row: images, hasVideo, promotion, deliveryEstimate, shippingCostText, readyToShip, easyReturn, productCertifications, attributes, supplier.inspectedBy, supplier.companyCertifications, supplier.staff, supplier.exportCountries, supplier.storeUrl, keyword, page, position, sortedBy and scrapedAt. With includeDetails, a details object adds categoryPath, attributes, variants, variantPrices, customization, sample, leadTimes, packaging, shipsFrom, certifications and the page's supplier facts.

How to read the price fields

• unitPriceAtMoq is the price you pay per unit when you order the minimum. On 76 of 76 laddered or single priced offers checked, it equals the product page's price at the MOQ.
• priceBasis: "price_range" means the price depends on the variant (colour, size, capacity): unitPriceAtMoq stays empty and unitPriceAtMoqMin and unitPriceAtMoqMax give the range. A single figure there would be a guess.
• isPerPiece is true only for pieces and units. Sets, pairs and bags are items too (unitClass: "item"); meters, kilograms and cartons are not.
• placeholderSuspected marks an MOQ of 100,000 or more, or a price of USD 0.01 or less, which suppliers often use as placeholders.
• outOfRange (with a price range set) is true when the price at MOQ lies outside your range.

The price summary row

One keywordSummary row per search, in USD per unit at the minimum order, over the offers this run delivered for that search (ads included unless excludeAds).

The headline is medianUsd, with lowerQuartileUsd, upperQuartileUsd, pricedOffers and headlineFigure naming where it comes from: the median price at MOQ of offers priced per item (pieces, units, sets, pairs and bags each as one sellable item) whose title holds every word of the keyword (perItemAllWords) when there are 5 or more of them; otherwise the first offers figure (firstOffers): the median price at MOQ of the first 10 offers priced per item in Alibaba's best match order (summaryOffers sets 5 to 50), counting only offers inside your price range, a variant priced offer at the low end of its range. With fewer than 5 of either, the headline stays empty and headlineNote says why. Beside it, unchanged: perPiece (per piece rows whose title carries at least half the keyword), perItem (the same over item units), perItemAllWords, firstOffers, byUnit and perPieceAllTitles. Variant priced rows, placeholders and rows outside your price range stay out of the other figures, counted in leftOutOfFigures.

How to get a typical factory price for a product idea: search the words a buyer would type and give a price range around the price you expect (minPrice, maxPrice, wide: for example a tenth to four times it). The range keeps accessories, spare parts and bulk listings out of every figure. The headline then reads, at the minimum order, the offers whose title holds every word of your keyword when there are 5 or more of them, else the first offers Alibaba ranks best inside your range.

How close it gets, measured honestly: on 9 product ideas checked by hand against product pages (saved answers, no range), the headline landed within 15 percent of the hand checked price at MOQ on 7 of 9 (8 of 9 against every relevant offer of the first 20). On 6 longer product ideas with known supplier prices and a price range each (real runs, 2026-11-06), it landed within 15 percent on 2 of 6, with a median error of 20 percent: long keywords rarely have 5 titles holding every word, so the first offers figure carries them, and a product sold mostly as a set or with accessories misses by about half. The low end of offers' ranges (the bulk price) missed by up to 60 percent.

Summary row of the prefilled search (yoga mat, 20 offers, read 2026-11-06, trimmed):

{
  "type": "keywordSummary",
  "keyword": "yoga mat",
  "currency": "USD",
  "headlineFigure": "perItemAllWords",
  "medianUsd": 7.84,
  "lowerQuartileUsd": 3.9925,
  "upperQuartileUsd": 13.725,
  "pricedOffers": 14,
  "headlineNote": null,
  "offersDelivered": 20,
  "perPieceOffers": 19,
  "adsCounted": 10,
  "perItemAllWords": { "count": 14, "median": 7.84, "lowerQuartile": 3.9925, "upperQuartile": 13.725, "min": 2.9, "max": 408 },
  "firstOffers": { "count": 10, "median": 4, "lowerQuartile": 3.715, "upperQuartile": 8.915, "min": 2.8, "max": 14.5, "firstN": 10 },
  "perItem": { "count": 14, "median": 7.84, "lowerQuartile": 3.9925, "upperQuartile": 13.725, "min": 2.9, "max": 408 },
  "perPiece": { "count": 14, "median": 7.84, "lowerQuartile": 3.9925, "upperQuartile": 13.725, "min": 2.9, "max": 408 },
  "complete": false,
  "stoppedBy": "maxResultsPerKeyword"
}

How much does it cost to scrape Alibaba?

You pay per event (Apify's pay per event model): per offer row delivered, once per search that delivered offers, and for the extras you turn on. Prices per 1,000 events by Apify plan:

Event	What it is	Free	Bronze	Silver	Gold
Offer	one offer row	$0.69	$0.63	$0.57	$0.49
Search	one search (a keyword, link or category, per supplier country) that delivered at least one billed offer, with its price summary row	$1.40	$1.28	$1.16	$1.00
Product detail	one product page read and delivered (includeDetails, or a product link)	$2.66	$2.43	$2.20	$1.90
Delivery country	once per run that reads as a buyer in a chosen country (shipToCountry or a residential proxy)	$28.00	$25.60	$23.20	$20.00

Platinum and Diamond plans pay the Gold price. The Actor start event is $0.0075 per run at the default 256 MB, at every plan (Apify charges it per GB of the run's memory). The Store pricing card shows these same prices per 1,000 events: "$0.49 / 1,000" on the Offer row means one offer costs 0.049 cents at Gold, and "$20.00 / 1,000" on the Delivery country row means one delivery country costs 2 cents at Gold.

Worked example on the Gold plan: one keyword with 20 offers (the prefilled run) costs $0.0075 start fee + 20 x $0.00049 + 1 x $0.001 = $0.0183. On the Bronze plan, 10 keywords of 100 offers each cost $0.0075 + 1,000 x $0.00063 + 10 x $0.00128 = about $0.65; with includeDetails add 1,000 x $0.00243 = $2.43.

Never billed: repeats of an offer already delivered in the run, offers your filters or excluded ids leave out, ads left out, a search that matches nothing, product pages that cannot be read, and the coverage, summary and status rows. A run refused for its input reads nothing from Alibaba. Only the run's start fee is charged. maxItems (1,000 by default) caps the paid rows of a run, and a row whose events do not fit your run's charge limit is held back, not delivered unpaid. Work that bills nothing has an allowance: a run whose searches match nothing stops after about 30 such searches at 256 MB (about 9 at 1 GB), with a free row that says so, and the searches after that point are not read. A run also holds at most 2,500 offers it has read but not yet delivered (many sorted or paused searches at once); past that, the newest are dropped, not charged, and a free row and the coverage row say so. Product pages (includeDetails, product links) are read one at a time about 6 seconds apart, about 500 to 600 an hour: a run that needs more than that needs a timeout above the default hour.

Input

At least one of keywords, searchUrls, productUrls or categoryIds. Every setting is described in the input schema; "Alibaba's filter" means it is sent to Alibaba, "on each row" means every delivered row is checked, and rows that fail are left out, free, and counted by reason on the coverage row.

What to search

Setting	What it does
keywords	product words, one search each; duplicates are read once; a pasted Alibaba link is read as a link
searchUrls	trade/search, search/page or catalog/<name>_cid<id> links, and showroom links (products/<words>.html, read as a search of their words, which the coverage row says); their words, category, page and filters apply, and a setting you give replaces the link's own; a link with a parameter this Actor does not read is refused with a free row naming it
productUrls	product links or ids, each read from its product page and billed as one Product detail
categoryIds	with keywords, each keyword is searched in each category; alone, each category is browsed (Alibaba's filter; the category is judged by Alibaba only)

A run takes at most 200 searches (keywords x categories x supplier countries, plus links) and 1,000 product links.

How many and in what order

Setting	What it does
maxResultsPerKeyword	offers per search, 1 to 960 (default 20); Alibaba serves one search 20 pages of 48 offers
maxPagesPerKeyword	result pages per search, 1 to 20 (default 20 in Alibaba's orders, 1 with an in-run sort)
summaryOffers	how many first offers the summary row's first offers figure counts, 5 to 50 (10 when left out)
startPage	the page each search starts at, 1 to 20
maxItems	paid rows of the whole run, 1 to 100,000 (default 1,000)
sortBy	best_match (default) and sales are Alibaba's own orders; price_low, price_high and response_rate are in-run sorts: Alibaba serves no such order, so the Actor sorts the offers it read and each row's sortedBy says so

Price and order quantity

Setting	What it does
minPrice, maxPrice	Alibaba's price filter in USD per unit; each row gets outOfRange, and a row whose whole range misses yours, or that states no price, is left out
maxMoq	Alibaba's filter, and on each row: only offers you can order in this quantity or less
minMoq	on each row: only offers whose MOQ is at least this
orderQuantity	adds unitPriceAtQuantity from each offer's ladder (empty, with belowMoq: true, below the offer's MOQ)

Supplier

Setting	What it does
supplierCountries	Alibaba's country filter, one search per country, and on each row (up to 20 countries)
verifiedOnly, verifiedProOnly	Alibaba's Verified and Verified Pro filters, and on each row
tradeAssuranceOnly	on each row
businessTypes	on each row: manufacturer, trading_company, distributor, buying_office, other (with verifiedOnly, manufacturer gives verified manufacturers)
minSupplierYears	on each row
minSupplierRating	on each row; 4.5 and 5 also use Alibaba's filter
companyCertifications	Alibaba's filter with the first name, and on each row for any of them (ISO, BSCI, IATF 16949)
excludeSupplierIds	supplier ids left out, free, never counted against a cap

Product

Setting	What it does
productCertifications	Alibaba's filter with the first name, and on each row for any of them (CE, ROHS, FCC, UL)
readyToShip	Alibaba's Ready to Ship filter, and on each row
samplesAvailable	Alibaba's samples filter (not checked on the row, which carries no sample flag); ads are left out under it
productTagId	Alibaba's product tag filter, and on each row
maxLeadTimeDays	on each row, from the search card's lead time
excludeAds	on each row: paid placements left out, free
onlyTitleMatches	on each row: offers whose title holds fewer than half of the keyword's words are left out, free
excludeOfferIds	offer ids of earlier runs left out, free

Product details and delivery

Setting	What it does
includeDetails	reads each delivered offer's product page, one at a time about 6 seconds apart, billed as one Product detail per page read
shipsFromCountries	on the product page: only offers that list one of these places of dispatch (needs includeDetails)
shipToCountry	reads Alibaba as a buyer in this country (71 countries) through an Apify residential exit there: delivery dates and shipping costs are that country's, prices stay in USD; billed once per run as Delivery country. Each result page is then read through a residential exit, which costs far more than a normal read: keep maxResultsPerKeyword at 20 or more per keyword, or the run stops early with a free row that says so
includePersonalData	with details, adds the supplier's contact person as the product page shows them; off by default

Advanced

includeRawOffer (Alibaba's own offer object under raw), extendOutputFunction (your JavaScript function runs on each row in a separate sealed process with no access to the run's secrets, files or the network, and its result is stored under custom), maxConcurrency (1 to 3 result pages in flight, default 2), maxRequestRetries (0 to 5, default 2), maxRunSeconds, debugLog and proxyConfiguration (left out: the run's own address with one Apify datacenter session behind it; or Apify datacenter from the start, or Apify residential with a country).

Example: verified manufacturers in China or Vietnam, a largest MOQ of 100, priced at 200 pieces:

{
  "keywords": ["silicone spatula set"],
  "supplierCountries": ["CN", "VN"],
  "verifiedOnly": true,
  "businessTypes": ["manufacturer"],
  "maxMoq": 100,
  "orderQuantity": 200,
  "maxResultsPerKeyword": 50
}

FAQ

How does Alibaba pricing work (price ranges and quantity tiers)?

An offer shows a price range and, on 67.9 percent of offers, a quantity ladder of 2 or more tiers whose unit price falls as the order grows, so the low end of the range is the price at the largest order. Each row carries the ladder in USD (priceTiers), the range (priceMin, priceMax) and the price at the minimum order (unitPriceAtMoq). priceBasis says where that price came from: a ladder, a single price, or a range that depends on the variant (price_range, where unitPriceAtMoqMin and unitPriceAtMoqMax give the range).

Why is the listed price lower than what the supplier quotes?

The low end of an Alibaba price range is the bulk price at the largest order: on 9 product ideas checked by hand, it missed the price at the minimum order by up to 60 percent. This Actor reports the price at the minimum order (unitPriceAtMoq, equal to the product page's price at MOQ on 76 of 76 laddered or single priced offers checked) and, with orderQuantity, the price at your own quantity.

What does MOQ mean on Alibaba, and how do I find low MOQ suppliers?

MOQ is the minimum order quantity a supplier sets for an offer; every row carries it as a number with its unit (moq, moqUnit, and Alibaba's text in moqText). Set maxMoq, for example 10, to keep only offers you can order in that quantity or less: Alibaba's own filter, checked again on each row.

How do I export Alibaba products to CSV or Excel?

Run the Actor with your keywords and open the dataset: the Offers view shows the offer columns (price at MOQ, unit, MOQ, supplier) and the Price summaries view the summary columns. Every row type is in the dataset, so keep the rows whose type is offer (or keywordSummary) in your spreadsheet, or ask the API for them. Both download as CSV, Excel or JSON from the run or through the API.

Does Alibaba have a product search API?

This Actor works as one: call automation_craft/alibaba-scraper through the Apify API with keywords, links or category ids and read the rows as JSON (examples below). It needs no Alibaba account or key, and reads Alibaba logged out.

Why does this Actor run with limited permissions?

Least privilege: it only reads and writes its own run's storages (the run's dataset and key value store). This Actor opens no named store, so it cannot read anything else in your account.

What this Actor does NOT do

• No Alibaba Guaranteed badge or filter: no Guaranteed flag is served on the search data or the product pages this Actor reads (0 of 15,185 saved search rows of 4,820 distinct offers, 0 of 139 product pages, and none of the 4,359 rows this Actor delivered in its own test runs).
• No standard or worldwide marketplace switch: the worldwide tab parameter changed nothing in the search data this Actor reads (3 tests).
• No store review lists and no supplier verification reports: both live on supplier pages under /company_profile/, which the supplier sites' robots.txt disallows. Rows carry the store rating, review count, service, shipping and product scores and review tags, the inspection company (supplier.inspectedBy), the company certifications and the profile link.
• No supplier search: this Actor delivers product offers, each with its supplier's facts, not supplier rows found by a supplier search.
• No buying requests (RFQ): those live on another Alibaba site, which this Actor does not read.
• No monitor of new or changed offers: each run delivers what it reads. To skip offers you already have, pass their ids in excludeOfferIds (free).
• No feedback note in the input: Apify gives a developer no access to other users' runs or inputs, so a note there cannot reach us. Please use the Issues tab.
• The summary statistics are over the offers this run delivered for a search, not over the whole of Alibaba's market for the keyword.
• Prices stay in USD, as Alibaba serves them: no conversion. A row Alibaba serves in another currency carries no USD figures (currency empty), never a converted one.
• Alibaba's order of page 1 changes between calls (5 calls in 77 seconds gave 5 orders), so two runs of one keyword deliver overlapping but not equal offers. totalReported on the coverage row is Alibaba's own count and moves between calls too.
• No more than 20 pages (960 offers) per search: pages 21 to 24 bring few or no new offers, and page 21 can re-serve pages 1 to 20 in one oversized answer.
• No solving of Alibaba's slider verification page. When Alibaba shows it, that door closes for the rest of the run, and the searches or products it leaves unread get free rows (one move to a fresh address is made only when the page comes within a door's first 3 calls).
• No checking of the category or the samples filter on the row: Alibaba alone applies them.
• No use of your own proxy links.
• Personal data is off by default: the supplier's contact person (name, gender, job title, photo) is empty unless you turn includePersonalData on. You are responsible for a lawful use of personal data, GDPR included.

API

curl -X POST "https://api.apify.com/v2/acts/automation_craft~alibaba-scraper/run-sync-get-dataset-items" \
  -H "Authorization: Bearer YOUR_APIFY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"keywords": ["yoga mat"], "maxResultsPerKeyword": 20}'

import { ApifyClient } from 'apify-client';
const client = new ApifyClient({ token: 'YOUR_APIFY_TOKEN' });
const run = await client.actor('automation_craft/alibaba-scraper').call({ keywords: ['yoga mat'], maxResultsPerKeyword: 20 });
const { items } = await client.dataset(run.defaultDatasetId).listItems();
const offers = items.filter((r) => r.type === 'offer');

from apify_client import ApifyClient
client = ApifyClient("YOUR_APIFY_TOKEN")
run = client.actor("automation_craft/alibaba-scraper").call(run_input={"keywords": ["yoga mat"], "maxResultsPerKeyword": 20})
items = client.dataset(run["defaultDatasetId"]).list_items().items
summary = [r for r in items if r["type"] == "keywordSummary"]

Changelog

• 0.1 (2026-11): first release. Offers by keyword, search link, category or product link with the USD price per unit at the minimum order, the quantity price ladder, MOQ and unit, supplier facts and the ad flag; a price summary row per search; the Product detail add-on. The full list is on the Changelog tab.

More data tools by Automation Craft

• IndiaMART Scraper: B2B Products, Prices, Suppliers
• AliExpress Search Scraper - Products and Prices
• Amazon Product Scraper - Search, Best Sellers
• Amazon Sponsored Products Scraper - Ads by Keyword
• Walmart Scraper: Products, Prices, Sellers
• Etsy Scraper: Listings, Prices, Ads, Reviews