Skip to content

Taobao Scraper for Product and Image Search

A managed scraping endpoint for Taobao pages that handles proxy rotation and JavaScript rendering behind a single HTTP request.

OUTPUT FORMATSHTMLJSONMarkdown

This page explains what you can pull from public Taobao pages, how the request path works, and how to estimate credit cost before you commit to a batch.

What the endpoint handles for you

Taobao pages are JavaScript-heavy and frequently guarded by IP-based blocking. Rather than assembling a headless browser fleet and a proxy pool yourself, you send a GET request with a Taobao URL on taobao.com and receive rendered HTML back, ready to parse. Proxy management and JavaScript rendering are part of the request path, not separate infrastructure you operate.

The practical effect is that your code focuses on two things: which URLs to fetch, and what to do with the returned markup. Everything between those two points — browser launch, rendering wait, proxy selection — is handled upstream.

Three workflows this fits

Most Taobao scraping work falls into one of three shapes. Knowing which one you are building determines which options you need.

Price tracking. Pull product prices and historical price changes per SKU to feed China-market pricing intelligence and competitive merchandising. This is a per-product, repeated-fetch pattern: the same SKU URLs on a schedule.

Seller storefront monitoring. Aggregate seller-level data — storefront name, total listings, rating — for marketplace-competition research. This is a per-seller pattern and stays on public storefront data; it is not a route for harvesting seller contact details.

Review mining. Pull customer reviews, star ratings, and review text per SKU to feed sentiment and product-feedback workflows. This is a paginated pattern, so request volume per SKU is higher than price tracking.

Fields available on a Taobao product page

A Taobao product page returns SKU, brand, price, seller storefront, ratings, review count, category, and the full image gallery. That set covers the three workflows above without additional parsing layers.

A useful field-to-workflow map:

  • SKU — the join key for price history and review history. Without a stable SKU, repeated fetches cannot be compared over time.
  • Price — the primary signal for pricing intelligence; pair it with a fetch timestamp you control.
  • Seller storefront — links a product to a seller record for competition research.
  • Ratings and review count — the aggregate layer; use review text extraction when you need the underlying comments.
  • Category — the grouping key for cross-product comparison.
  • Image gallery — useful for catalog and merchandising checks.

The request path, step by step

  1. Create an account. You get an API key and 1,000 free credits to start.
  2. Install a client or skip it. SDKs exist for Python, Node.js, Go, PHP, Ruby, and Java; plain cURL works too. Each Taobao scrape becomes a single function call.
  3. Send a GET request with your API key and a Taobao URL on taobao.com. JavaScript renders, residential proxies rotate, and rendered HTML comes back.
  4. Add options only when the page needs them. render_js for JavaScript, stealth_proxy for CAPTCHAs, country_code for geo-targeting, ai_extract_rules for natural-language extraction.

Choosing how to extract

There are two extraction routes and they fail differently, which matters when you are debugging.

Selector-based extraction uses extract_rules with CSS selectors (or XPath). It is deterministic: the same selector against the same markup returns the same result. When Taobao changes its markup, the selector silently returns nothing rather than erroring — so validate that each rule produced a non-empty value before writing a row.

AI-driven extraction uses ai_extract_rules with a natural-language specification and returns structured JSON without selectors. This is more forgiving of markup changes but is a different contract: you describe the fields you want rather than pointing at elements.

Both routes return structured JSON. Pick selectors when you need stable, auditable output; pick natural-language rules when the page structure shifts often or you are prototyping.

Estimating credit cost before a batch

Credits are consumed per successful request, and options stack. The per-request figures are: 5 credits for JS rendering, 25 for premium_proxy with JS rendering, and 75 for stealth_proxy. Separately, the pricing tooltip notes that premium proxies cost 10 or 25 credits depending on whether JS rendering is disabled or enabled.

A worked example: a 1,000-SKU price-tracking run that needs JavaScript rendering costs 1,000 × 5 = 5,000 credits. The same run with premium proxies and JS rendering costs 1,000 × 25 = 25,000 credits. If a portion of those pages hit CAPTCHAs and you escalate to stealth_proxy, those requests cost 75 credits each instead.

The planning consequence: decide your option tier per workflow before scaling, because the multiplier between the cheapest and most expensive path is large. Start with the 1,000 free credits on a representative sample of your target pages, measure how many need escalation, then extrapolate.

Pre-flight checklist and failure distinctions

Before a production batch, confirm:

  • [ ] Every target URL is a public page on taobao.com, not post-login content.
  • [ ] Your extraction rules return non-empty values on a sample of at least a few page variants.
  • [ ] You know which option tier each workflow needs, and have costed it.
  • [ ] You record a fetch timestamp alongside each row so price and review history is comparable.
  • [ ] You have reviewed the target site's terms of service for your jurisdiction and use case.

Three failure modes look similar but need different fixes:

  • Blocked request — you get an error or a challenge page instead of product markup. Fix: escalate to stealth_proxy or adjust geo-targeting.
  • Empty extraction — the request succeeded but your selector matched nothing. Fix: the page markup changed; update the selector.
  • Partial extraction — some fields populated, others empty. Fix: usually a selector scoped too broadly or too narrowly; check each rule independently.

Plan tiers and what changes between them

Paid plans start at $19/month and scale through higher credit allowances and concurrency. Across the standard tiers, JavaScript rendering, rotating and premium proxies, geotargeting, and screenshots with extraction rules are included. Priority email support, a dedicated account manager, and team management appear at the Business tier and above.

All prices are exclusive of VAT. Credits are valid for one month and do not roll over, and requests start at 1 credit each before options are applied. The plan table below lists every tier with its credit allowance and concurrency figure.

Plan tiers, credits, and concurrency
PlanPrice per monthAPI creditsConcurrent requestsPriority supportAccount managerTeam management
Hobby$1975,00025Not includedNot includedNot included
Freelance$49250,00050Not includedNot includedNot included
Startup (Recommended)$991,000,000100Not includedNot includedNot included
Business$2493,000,000200IncludedIncludedIncluded
Business +$5998,000,000400IncludedIncludedIncluded
Enterprise 14$99914,000,000500IncludedIncludedIncluded
Enterprise 24$1,59924,000,000600IncludedIncludedIncluded
Enterprise 41 (Recommended)$2,39941,000,000700IncludedIncludedIncluded
Enterprise 69$3,59969,000,000800IncludedIncludedIncluded
Enterprise 120$5,799120,000,000900IncludedIncludedIncluded

Prices are exclusive of VAT. Credits are valid for one month and do not roll over. JavaScript rendering, rotating and premium proxies, geotargeting, and screenshots with extraction rules are included on every tier listed.

Frequently asked questions

Do I need a Taobao account to scrape it?

No external account is required. You sign up for an API key, which comes with 1,000 free credits and no card, then send the URL and the rendering and proxy handling is done for you. This applies to public pages only; post-login content is out of scope.

Is scraping Taobao legal?

Public-facing pages can usually be scraped, but you should review the target site's terms of service before deploying and avoid post-login content. Legality depends on your jurisdiction and use case, so treat this as a starting point rather than legal advice.

What data can I extract from Taobao?

Whatever is visible on the public page. You have two ways to shape the output: point extract_rules at specific elements with CSS selectors, or describe the fields you want in plain language through ai_extract_rules and let the AI route assemble them. Either way the response is structured JSON. On product pages that typically covers SKU, brand, price, seller storefront, ratings, review count, category, and the image gallery.

How do I start scraping Taobao?

Sign up, grab your API key, and send a GET request with the target URL. The page is rendered, proxies rotate, and data comes back. You get 1,000 free credits to test with.

Which programming languages can I use?

Any. SDKs are published for Python, Node.js, Go, PHP, Ruby, and Java, and plain cURL works too. Pick whatever your stack already uses.

How much does it cost to scrape Taobao?

You get 1,000 free credits on signup with no card required. Paid plans start at $19/month for higher volume and concurrency. Credit cost per request depends on options: 5 credits for JS rendering, 25 for premium_proxy with JS, and 75 for stealth_proxy.

What is the difference between a blocked request and an empty extraction?

A blocked request returns an error or a challenge page instead of product markup, and is usually fixed by escalating proxies or adjusting geo-targeting. An empty extraction means the request succeeded but your selector matched nothing, which usually means the page markup changed and the selector needs updating. Checking whether you received product markup at all is the fastest way to tell them apart.