Alibaba Scraper — Verified Supplier Ranking

Alibaba Scraper — Products, Supplier Trust Scores and Reorder Rates

Alibaba Scraper turns any public Alibaba.com search or category page into a ranked supplier report: every product row carries a derived 0–100 supplier trust score with a full component breakdown, plus the supplier's reorder rate — computed locally from the same listing payload the scraper already downloads, no extra requests and no login. Unlike scraping frameworks that hand you raw HTML, Alibaba Scraper returns typed JSON — ready for your sourcing spreadsheet, your database, or your LLM context window without any parsing. This guide covers every input and output field, the exact trust-score formula, and how sourcing, research, and monitoring teams deploy the Actor in production.

🧭 What does Alibaba Scraper do?

Alibaba Scraper fetches one or more Alibaba search-result or category-browse URLs and extracts every product/supplier offer on the page, without requiring an Alibaba account, login, or cookie export — the Actor sends unauthenticated requests with Chrome-impersonated TLS fingerprints. For each offer it returns the full base product record (title, price, MOQ, supplier identity, images, review and shipping scores, badges, certifications) plus two deliverables no base Alibaba scraper computes: a weighted supplier trust score and a harvested reorder rate.

• Scrapes any logged-out Alibaba search-result or category-browse URL, including one already filtered by keyword, category, or price on Alibaba's own UI
• Computes supplierTrustScore (0–100) from verification status, gold-supplier years, review score, and transaction volume — with a full per-component supplierTrustBreakdown
• Harvests reorderRate / reorderRatePct from the offer's selling-point badges, when Alibaba surfaces one
• Optionally ranks the saved dataset by trust score, highest first (sortByTrust)
• Preserves every field the base product scrape returns — 38 raw offer fields on top of the 9 derived ones
• Auto-derives how many result pages to open per URL from your row target, or accepts a manual page cap
• Escalates through a residential-proxy retry ladder when a fetch looks blocked, so runs keep going without you managing sessions

⚡ Features & capabilities

Alibaba Scraper's feature set splits into three layers: what it extracts, how it survives Alibaba's anti-bot defenses, and how it stacks up against other Alibaba tools on the Apify Store.

Core output fields on every row (exact names from the source): title, price, moq, moqV2, companyName, countryCode, goldSupplierYears, reviewScore, reviewCount, soldOrder, badges, certifications, productUrl, mainImage, multiImage, plus the derived supplierTrustScore, supplierTrustBreakdown, reorderRate, and reorderRatePct. Fetching uses curl_cffi with a chrome120 TLS/HTTP fingerprint, and every request escalates through direct fetch → Apify Proxy → Apify Residential Proxy with sticky sessions if the previous attempt looks blocked.

A response "looks blocked" and triggers the next rung of the escalation ladder when the HTTP status isn't 200, the returned HTML is under 800 characters, or the page is missing the window.__page__data_sse / _offer_list markers the Actor parses. Retries between attempts back off exponentially — delay = min(2.5, 0.6 × 1.6^attempt) seconds, plus a small random jitter — and once a residential session successfully returns a page, that session ID is reused (sticky) for subsequent requests instead of rotating on every call.

How Alibaba Scraper compares to other Alibaba scrapers

Feature	Alibaba Scraper (this Actor)	zen-studio/alibaba-scraper	devcake/alibaba-supplier-scraper
Computed trust/credibility score	Yes — 0–100 composite with auditable per-component breakdown	Not documented (returns raw reorderRate/badge fields, no computed score)	Not documented (returns raw verification flags, no computed score)
Reorder rate	Yes — parsed from selling-point badges (reorderRate, reorderRatePct)	Yes — raw sellingPoints/reorderRate text field	Not documented in their field list
Pricing model	Pay-per-event, charged per row_result	Pay-per-event — $4.99 per 1,000 products (as observed on their listing, July 2026)	Pay-per-event — $0.01 Actor Start + $0.005 per supplier (as observed on their listing, July 2026)
Account/login required	No	Not documented as required; their README states "no account, login, or cookie export is involved" (as observed, July 2026)	Not documented
Anti-bot handling	curl_cffi Chrome TLS impersonation + Apify Residential Proxy escalation ladder with sticky sessions (in source)	"Residential proxy is handled automatically" (as stated on their listing, July 2026)	Not documented
Full base product schema preserved	Yes — 38 raw offer fields plus 9 derived fields on every row	Documented as "100+ fields" including supplier company profile, SKU matrix, certifications (as claimed on their listing, July 2026)	~14 named fields across verification and factory-intelligence groups (as listed in their README, July 2026)

If your use case is feeding structured supplier data to an LLM or a scoring pipeline, the trust-score row is the decision-maker — re-implementing credibility scoring inside your own agent loop from raw badge fields is a reliability failure mode, not a shortcut.

When another tool might suit you better

If you need Alibaba's deepest supplier due-diligence data — the full SKU attribute matrix, ladder pricing tiers, an SGS-style verification report, or a factory-tour video — zen-studio/alibaba-scraper documents a materially larger schema (100+ fields, per their listing) that this Actor does not attempt to replicate. This Actor is built for one job: turning a listing page into a scored, sortable shortlist. If you only need a bare supplier list at the lowest possible cost and prefer flat monthly pricing over pay-per-event, a simpler tool such as shareze001/scrape-alibaba-suppliers (flat-price plan, per their listing) may be cheaper for light, occasional use.

Why do developers and data teams scrape Alibaba?

Alibaba is the largest B2B sourcing marketplace most import-driven businesses touch first, and its front end changes often enough that manual browsing doesn't scale past a handful of SKUs.

🏢 B2B sourcing and procurement teams

Sourcing teams paste category or keyword URLs, set maxSuppliers to the shortlist size they need, and turn on sortByTrust so the highest-scored suppliers land at the top of the dataset before anyone opens a single product page. The team then filters the exported rows on trustVerified, trustGoldYears, and reorderRatePct to build a call list, instead of manually reading gold-supplier badges across dozens of open tabs.

📊 AI training data and RAG indexing

title, loopSellingPoints, pcLoopSellingPoints, and the certification/badge arrays are the high-information text fields worth indexing for retrieval — they carry Alibaba's own marketing language and structured claims about a product or supplier. For training data, supplierTrustScore, trustGoldYears, trustReviewScore, and trustTransactions are numeric fields with consistent structure across every row, useful as labeled features without any additional normalization.

📱 Competitive and market intelligence

Track price, soldOrder, and reorderRatePct for the same productId across scheduled runs to see which suppliers are gaining order volume or dropping price, and which listings are picking up a reorder-rate badge they didn't carry before.

🔬 Research and academic use

Public listing data — pricing distributions, gold-supplier tenure, review-score spread across a category — supports market-structure and B2B e-commerce research. The Actor returns only what Alibaba serves to a logged-out visitor; it does not access private supplier communications or buyer records.

🎥 Product and SaaS development

Supplier-directory products, sourcing-recommendation tools, and vendor-monitoring dashboards can be built directly on the dataset schema — supplierTrustScore and supplierTrustBreakdown give a ready-made ranking signal without your product having to invent and maintain its own scoring model.

🍚 Input parameters

All fields are read from .actor/actor.json. Only supplierListingUrls is required.

Parameter	Required	Type	Default	Constraints	Description
supplierListingUrls	Yes	array (string list)	—	one URL per line	Alibaba search-result or category-browse URLs to build the report from. Works with any listing view you can open logged-out in a browser. Paste many URLs to combine several searches into one report. Backward-compatible: the base urls key is also accepted.
maxSuppliers	No	integer	10	min 1, max 100000	Total number of product/supplier rows to include in the report, across all URLs. Alibaba returns roughly 48 products per result page; the Actor opens just enough pages to reach this total and trims any overflow. Backward-compatible: the base maxItems key is also accepted.
sortByTrust	No	boolean	false	—	When true, collected rows are ranked by supplierTrustScore (highest first) before being saved; rows with no scoreable signal sort last. When false (default), rows save live in page order as each page finishes.
maxPages	No	integer	0	min 0, max 5000	Optional hard cap on result pages opened per URL. 0 (default) lets the Actor derive the page count automatically from maxSuppliers. A positive value opens exactly that many pages per URL regardless of the row target.
proxyConfiguration	No	object (proxy editor)	{"useApifyProxy": true, "apifyProxyGroups": ["RESIDENTIAL"]}	—	Apify Proxy configuration. Residential is enabled by default; Alibaba blocks datacenter IPs aggressively, so keeping Residential on is recommended.

No parameter in this schema is a credential or API key — the Actor needs no Alibaba login. The only credential-like setting is optional and lives outside the input schema: setting the CUSTOM_PROXY_URL environment variable (format http://user:pass@host:port) under your Actor's Environment variables in Apify Console routes requests through your own residential proxy before Apify Proxy is tried.

Example input:

{
  "supplierListingUrls": [
    "https://www.alibaba.com/trade/search?fsb=y&IndexArea=product_en&categoryId=127734135&has4Tab=true&keywords=Men%27s+Jackets&originKeywords=Men%27s+Jackets&tab=all&page=1"
  ],
  "maxSuppliers": 50,
  "sortByTrust": true,
  "maxPages": 0,
  "proxyConfiguration": {
    "useApifyProxy": true,
    "apifyProxyGroups": ["RESIDENTIAL"]
  }
}

Supported URL types and input formats

supplierListingUrls accepts two input shapes per entry: a plain string, or an object with a url key ({ "url": "https://..." }) — both are normalized internally. Any Alibaba search-result or category-browse page that renders without logging in will work, because the Actor parses the same window.__page__data_sse*._offer_list JSON blob the page embeds for its own product grid.

• Keyword search, single page: https://www.alibaba.com/trade/search?keywords=led+lights&page=1
• Keyword search filtered to a category: the same search URL with categoryId and has4Tab=true appended, exactly as Alibaba's own filter UI produces it — set your filters on Alibaba first, then copy the address bar
• Multiple sources combined into one report: list several search or category URLs, one per line; maxSuppliers is distributed across all of them and the resulting rows are combined into a single dataset

How page count auto-derives from maxSuppliers. When maxPages is left at its default of 0, the Actor computes, per source URL: per_url_target = ceil(maxSuppliers / number_of_urls), then estimated_pages = ceil(per_url_target / 48) (48 is the Actor's hint for products per Alibaba result page), then opens min(5000, estimated_pages + 1) pages for that URL — the +1 covers pages that return fewer than 48 usable offers. Setting maxPages to a positive value overrides this and opens exactly that many pages per URL regardless of maxSuppliers.

📦 Output format

Every collected offer is pushed to the Apify dataset as one flat JSON object — no nested product/supplier split. Each row is the full 38-field base offer plus 9 derived trust and reorder fields, for 47 fields total. The dataset's default view shows 17 of those 47 as table columns; every other field is still present on the row and available via JSON, CSV, XML, Excel, or the Apify API export.

Output for a product/supplier row

{
  "badges": [],
  "certifications": [
    { "prefixIcons": [{ "name": "ISO9001", "url": "https://s.alicdn.com/@sc04/kf/iso9001.png" }] }
  ],
  "chatToken": "RWRIaHZYd0dEaGQxU2FpNitNWmRq",
  "companyId": "238181772",
  "companyLogo": "https://sc02.alicdn.com/kf/H96f60213b26c423b840d3b928e1a3772S.jpg",
  "companyName": "Shenzhen Witop Technology Co., Ltd.",
  "contactSupplier": "//message.alibaba.com/msgsend/contact.htm?action=contact_action",
  "countryCode": "CN",
  "customGroup": "8",
  "displayStarLevel": "2",
  "eurl": "https://us-click.alibaba.com/ci_bb?ot=local&a=1403902656",
  "goldSupplierYears": "10",
  "id": "1601386544983",
  "isShowAd": false,
  "loopSellingPoints": [
    { "loopType": "re_order_rate", "text": "26% reorder rate" }
  ],
  "lyb": false,
  "mainImage": "https://s.alicdn.com/@sc04/kf/Ha15c3c42c6774e72ad0d297857aca960u.jpg",
  "moq": "10 meters",
  "moqV2": "10",
  "multiImage": [
    "https://s.alicdn.com/@sc04/kf/Ha15c3c42c6774e72ad0d297857aca960u.jpg",
    "https://s.alicdn.com/@sc04/kf/H2b8e1f0c9a124e7d8f3.jpg"
  ],
  "pcLoopSellingPoints": [
    { "loopType": "re_order_rate", "text": "26% reorder rate" }
  ],
  "price": "$0.53-0.63",
  "productId": "1601386544983",
  "productScore": "4.6",
  "productUrl": "https://www.alibaba.com/product-detail/Good-Price-24V_1601386544983.html",
  "reviewCount": "19",
  "reviewScore": "4.5",
  "shippingScore": "4.5",
  "showAddToCart": true,
  "showCrown": false,
  "soldOrder": "3200",
  "supplierHomeHref": "https://witop-led.en.alibaba.com/",
  "supplierHref": "https://witop-led.en.alibaba.com/company_profile.html",
  "supplierService": "4.4",
  "supplierServiceScore": "4.4",
  "title": "Good Price 24V 32.8ft Flexible LED Tape Smart RGB 10M Full Kit LED Neon Strip",
  "tmlid": "8pctgRBMALNrPrdp7AiuUA==",
  "trackInfo": "dynamic_tag:pidgroup=pc;testgroup=comp",
  "supplierTrustScore": 78,
  "supplierTrustBreakdown": {
    "verified": { "value": true, "score": 100, "weight": 0.20 },
    "goldYears": { "value": 10, "score": 100, "weight": 0.20 },
    "reviewScore": { "value": 4.5, "score": 90, "weight": 0.30 },
    "responseRate": { "value": null, "score": null, "weight": 0.15 },
    "transactions": { "value": 3200, "score": 87, "weight": 0.15 }
  },
  "trustVerified": true,
  "trustGoldYears": 10,
  "trustReviewScore": 4.5,
  "trustResponseRate": null,
  "trustTransactions": 3200,
  "reorderRate": "26% reorder rate",
  "reorderRatePct": 26.0
}

How supplierTrustScore is computed. The composite is a weighted average of five sub-scores, each normalized to 0–100:

Component	Weight	Source field	Sub-score formula
verified	0.20	goldSupplierYears, badges, certifications	100 if goldSupplierYears > 0 or badges or certifications is non-empty, else 0
goldYears	0.20	goldSupplierYears	min(goldYears, 10) / 10 × 100
reviewScore	0.30	reviewScore	clamp(reviewScore, 0, 5) / 5 × 100
responseRate	0.15	not present in the search-listing payload	always null in this Actor — the field lives only on the supplier's own minisite, which this Actor does not fetch
transactions	0.15	soldOrder	min(log10(transactions + 1) / 4, 1) × 100

supplierTrustScore = round( Σ(sub-score × weight) / Σ(weight of the components that have a value) ). A component with no source value (always responseRate; occasionally goldYears, reviewScore, or transactions when Alibaba omits them) is dropped from both the numerator and the denominator, so the remaining weights renormalize and the score always reflects only real, present signals — never a faked stand-in. If every component is missing, supplierTrustScore is null. supplierTrustBreakdown exposes each component's raw value, its 0–100 score, and its weight, so the composite is fully auditable from the row alone.

How reorderRate is extracted. The Actor scans the offer's loopSellingPoints, then pcLoopSellingPoints, for an entry whose loopType contains re_order or reorder, or whose text matches re-?order\s*rate (case-insensitive). The matched entry's display text becomes reorderRate (e.g. "26% reorder rate"), and a \d+(?:\.\d+)?\s*% match inside that text becomes the numeric reorderRatePct. If no offer selling point carries a reorder signal, both fields are null — Alibaba does not surface a reorder rate on every listing.

Every dataset row is a single charged row_result event — there is no separate uncharged accounting or error row to filter out; the Actor only calls push_data once per collected offer.

Schema stability and export options

Field names are read directly from Alibaba's own _offer_list JSON payload and stay stable across the Actor's runs as long as Alibaba doesn't rename the underlying keys in that payload; fields Alibaba omits from a given offer are returned with their documented default (empty string, empty array, or false) rather than being dropped from the row. Export the dataset as JSON, CSV, Excel, XML, RSS, or HTML from the run's Storage tab in Apify Console, or pull it programmatically through the Apify API or apify-client.

💡 Alibaba Scraper strategy guide

🎯 Strategy 1: Real-time enrichment pipeline

Trigger a run whenever a new sourcing request comes in: pass the category or keyword URL as supplierListingUrls, set maxSuppliers to your shortlist size, and turn on sortByTrust. Read the returned rows back through the Apify API and append supplierTrustScore, trustVerified, reorderRatePct, and companyName to the sourcing record in your PIM, spreadsheet, or CRM — the supplier's credibility signal lands on the same record as the product, in one API round trip.

🎯 Strategy 2: Scheduled monitoring and alerting

Set up an Apify Schedule to re-run the same supplierListingUrls on a cadence you control. Diff each new run's rows against the previous run's, keyed on productId, and alert when supplierTrustScore drops, price moves outside your threshold, or a reorderRate badge appears or disappears on a supplier you're tracking. This turns a one-off shortlist into an ongoing supplier-risk watch without any code beyond the diff step.

🎯 Strategy 3: Bulk dataset build

Feed a long list of category or keyword URLs into supplierListingUrls in one run, and set maxSuppliers as high as the 100000 schema maximum allows. The Actor's own concurrency is capped at 3 concurrent page fetches (MAX_CONCURRENCY = 3 in the source), so large runs progress through pages steadily rather than all at once. Export the finished dataset to CSV or pull it via the API to build a supplier scorecard or a category-wide pricing dataset.

Strategy comparison at a glance

Strategy	Best for	Run pattern	Output format
Real-time enrichment	Sourcing teams scoring inbound requests	On-demand run per request, via API	JSON via API, written back to your system
Scheduled monitoring	Watching known suppliers or categories over time	Apify Schedule, recurring	JSON diffed run-over-run
Bulk dataset build	Research or category-wide sourcing datasets	Single large run, many source URLs	CSV/JSON export from Storage or API

Related Alibaba scrapers & tools

Scrapio does not currently publish a second Alibaba-focused Actor, so there is no in-account sibling for supplier reviews or the 1688 domestic marketplace. For adjacent marketplace data on other platforms, these Scrapio Actors cover the same product/seller entity type — find them under the Scrapio account on the Apify Store:

Scraper	What it extracts
Amazon Search Products Scraper	Amazon search-result products by keyword
Amazon Product Scraper — Buy Box Seller Monitor	Amazon product listings with buy-box price and seller data
Amazon ASINs Scraper	Amazon product data by ASIN
Amazon Today's Deals Scraper	Amazon discount/deal listings by category and price

None of these cover Alibaba directly; if your workflow needs the same trust-score and reorder-rate treatment applied to a different B2B or wholesale marketplace, that would need a separate Actor built for that platform's payload shape.

How to integrate Alibaba Scraper with your stack

Alibaba Scraper works with any language or tool that can call the Apify API — there is no separate REST API of its own to learn, and no MCP server is documented for this Actor. Under the hood the Actor runs on the apify Python SDK (apify>=2.0.0) and curl-cffi (curl-cffi>=0.7.0) for the Chrome-impersonated fetch layer, packaged in the standard apify/actor-python Docker base image — none of that matters to how you call it, but it explains why the Actor's own retry and proxy-escalation behavior lives in the run rather than in a client library you'd need to install.

Python

from apify_client import ApifyClient
import csv

client = ApifyClient("<YOUR_APIFY_TOKEN>")

run_input = {
    "supplierListingUrls": [
        "https://www.alibaba.com/trade/search?keywords=led+lights&page=1",
    ],
    "maxSuppliers": 50,
    "sortByTrust": True,
}

run = client.actor("<YOUR_USERNAME>/alibaba-scraper-supplier-trust-reorder-report").call(
    run_input=run_input
)

rows = list(client.dataset(run["defaultDatasetId"]).iterate_items())

with open("alibaba_suppliers.csv", "w", newline="", encoding="utf-8") as f:
    writer = csv.DictWriter(f, fieldnames=[
        "title", "companyName", "supplierTrustScore", "reorderRatePct", "price"
    ])
    writer.writeheader()
    for row in rows:
        writer.writerow({
            "title": row.get("title"),
            "companyName": row.get("companyName"),
            "supplierTrustScore": row.get("supplierTrustScore"),
            "reorderRatePct": row.get("reorderRatePct"),
            "price": row.get("price"),
        })

print(f"Saved {len(rows)} supplier rows to alibaba_suppliers.csv")

Node.js

import { ApifyClient } from 'apify-client';

const client = new ApifyClient({ token: '<YOUR_APIFY_TOKEN>' });

const run = await client.actor('<YOUR_USERNAME>/alibaba-scraper-supplier-trust-reorder-report').call({
    supplierListingUrls: [
        'https://www.alibaba.com/trade/search?keywords=led+lights&page=1',
    ],
    maxSuppliers: 50,
    sortByTrust: true,
});

const { items } = await client.dataset(run.defaultDatasetId).listItems();

for (const row of items) {
    console.log(row.companyName, row.supplierTrustScore, row.reorderRatePct);
}

Async and scheduled pipelines

For large runs, start the Actor with client.actor(...).start() instead of .call() so your caller doesn't block, then poll the run status or subscribe to an Apify webhook on the ACTOR.RUN.SUCCEEDED event to pull the dataset once the run finishes. For recurring extraction, configure an Apify Schedule on the Actor directly in Apify Console rather than triggering runs from your own cron job.

Who needs Alibaba Scraper? (Use cases & industries)

🏢 B2B sourcing and procurement teams

Build a call list from sortByTrust-ranked rows, filtering on trustVerified and trustGoldYears before opening a single supplier's storefront — the scoring already happened during the run.

📊 AI/ML and data engineering teams

Feed supplierTrustScore, trustReviewScore, and trustTransactions into a downstream ranking model as pre-computed, auditable features, or index title and the selling-point text fields for a sourcing-assistant RAG pipeline.

📱 Import/export and dropshipping businesses

Pull price, moq, and reorderRatePct across a category to compare unit economics and reorder demand across suppliers before committing to a sample order.

🔬 Researchers

Public listing data supports B2B marketplace research — pricing distributions, gold-supplier tenure, and review-score spread by category — using only what Alibaba serves to a logged-out visitor.

Is it legal to scrape Alibaba?

Scraping publicly accessible web pages is generally lawful in most jurisdictions; the data Alibaba Scraper returns is limited to what Alibaba serves to any logged-out visitor, with no login or paywall bypassed. Running this Actor against Alibaba may still violate Alibaba's own Terms of Service, which is a contractual and civil matter between you and Alibaba, not a criminal one — review Alibaba's terms before using the data commercially, particularly for high-volume or automated collection. This Actor's output is business and product data — supplier company names, product listings, prices, review scores, gold-supplier tenure — not personal data about identifiable individuals: the source code does not extract a contact person's name, email, or phone number anywhere in the row it builds. Because no personal data is collected, GDPR and CCPA's personal-data obligations do not attach to the fields this Actor returns; ordinary contract and database-rights considerations around Alibaba's own listing content still apply. Alibaba Scraper returns only publicly accessible data. What you do with that data is your responsibility — consult legal counsel for commercial applications.

❓ Frequently asked questions

Does Alibaba Scraper work without an Alibaba account?

Yes. The Actor sends unauthenticated requests with an empty cookie jar and Chrome-impersonated headers — no Alibaba login, session, or cookie export is used or required anywhere in the source.

How does Alibaba Scraper handle Alibaba's anti-scraping measures?

It fetches with curl_cffi using a chrome120 TLS/HTTP fingerprint, and if a response looks blocked (non-200 status, a too-short page, or a missing _offer_list marker) it escalates: direct fetch, then Apify Proxy, then Apify Residential Proxy with up to 6 retries and sticky sessions, with exponential backoff between attempts (base 0.6s, capped at 2.5s).

Can I run Alibaba Scraper at scale without getting blocked?

maxSuppliers accepts up to 100,000 rows and maxPages up to 5,000 pages per URL, but the Actor caps itself to 3 concurrent page fetches at a time (MAX_CONCURRENCY = 3 in the source) rather than opening pages all at once. No uptime or block-rate guarantee is published; keeping Residential proxy enabled is the documented recommendation for reliability against Alibaba's datacenter-IP blocking.

How fresh is the data Alibaba Scraper returns?

Every run fetches live from Alibaba.com at request time — the Actor sends cache-control: max-age=0 on every request and does not cache or reuse HTML between runs.

Why is trustResponseRate always null?

Alibaba's search-listing payload doesn't carry a supplier response-rate figure — that number only appears on the supplier's own storefront/minisite page, which this Actor doesn't fetch. Rather than approximate it, the Actor always returns trustResponseRate: null and excludes it from the supplierTrustScore calculation entirely, renormalizing the remaining four components' weights instead.

What happens if my search URL returns zero products?

The run doesn't fail — the Actor logs "This page returned no products — moving on" for that page and continues to the next page or URL. If every source URL returns nothing, the final dataset simply has fewer rows than your maxSuppliers target, and the run still finishes successfully.

Which fields work best for AI training and RAG indexing?

For RAG, index title, loopSellingPoints/pcLoopSellingPoints text, and the certification/badge arrays — they carry Alibaba's own descriptive and credential language. For training features, supplierTrustScore, trustGoldYears, trustReviewScore, and trustTransactions are numeric and structurally consistent across every row, returned as typed primitives that need no normalization before use.

Does Alibaba Scraper work with Claude, ChatGPT, and other AI agent tools?

No MCP server is documented for this Actor. It is callable as a standard Apify Actor run by any agent framework that can call the Apify API — every response is typed JSON, so no HTML parsing step sits between the Actor's output and your agent's context window.

How does Alibaba Scraper compare to other Alibaba scrapers?

zen-studio/alibaba-scraper documents a deeper schema (100+ fields, per their listing as of July 2026) including SKU matrices and verification reports, but does not document a computed trust or credibility score. devcake/alibaba-supplier-scraper documents verification flags and factory-intelligence fields at $0.01 + $0.005/supplier (per their listing as of July 2026), also without a computed composite score. Alibaba Scraper's differentiator is the auditable supplierTrustScore/supplierTrustBreakdown and parsed reorderRate computed on top of the full base schema, in a single pay-per-row-result run.

Disclaimer

Alibaba Scraper extracts only publicly available data from Alibaba.com. This tool is intended for lawful use cases only. Users are responsible for complying with Alibaba's terms of service and applicable data protection laws in their jurisdiction.