One Year In: The Web Scraping Club Tested ScrapeUnblocker on Google, Shopee and Shein

ScrapeUnblocker has been on the market for about a year. In web scraping that is not long: the companies we are usually compared with have been selling proxies and scraping APIs for a decade or more.

So it meant a lot to us when Pierluigi Vinciguerra, the founder of The Web Scraping Club, one of the most widely read newsletters in our industry, agreed to test us. His review, “Testing ScrapeUnblocker on Google and Shopee”, is out now. We asked for the test. He picked the targets, wrote his own client, ran the calls and published what he saw, including the parts he did not like.

This post sums up what he tested, what he found, and what we changed in the days after. Read the full article for the details, the code and his own words.

Why These Three Targets

Pierluigi did not pick easy websites. He went for the ones the industry talks about when something stops working:

• Google, which keeps getting harder to scrape and now rewrites the result links it serves to clients it considers automated.
• Shopee, with custom bot protection, accounts and strict geographic controls. One of his own earlier Lab articles ended without a working Shopee scraper.
• Shein, about which he wrote:

Shein is another hard target, and when talking with other professionals, it’s sometimes not even economically viable to extract data from there

He treated ScrapeUnblocker as a black box. His client made one attempt per call, with no retries, so every failure would show up in the numbers. All calls ran from his laptop in Italy on October 25, 2026.

The Results

Target	Calls	Result	Latency
Google (/serpApi)	10	10 of 10 at the first attempt	3.8 to 8.1 s, median 6.3 s
Shopee plugin	10	10 of 10 at the first attempt	2.6 to 9.3 s, median 4.8 s
Shein plugin + product pages	10	10 of 10	1.7 to 6.7 s after a slow first call

In his words, all three targets worked on every request he sent: “32 calls without a single error”.

A few details from the review that matter to anyone building on this data:

• Google results with real URLs. Google now hides the destination of a result behind its own link for automated clients. “Out of the 89 organic results I collected, none pointed to a Google domain.” Every result came back with the real destination URL.
• AI Overviews included. The AI Overview was there on 8 of 10 queries, as text and Markdown, with the cited sources and references. For brand monitoring in AI answers, he called it probably the most useful part of the payload.
• Shopee across markets. Six searches in six markets returned 240 products in six currencies: “about 25 seconds of total waiting, which is quite a change” compared with his earlier attempt.
• Shein product data. Every search returned 20 products with price, discount, rating, review count and store, and the product pages came back as structured JSON with material, fit, care instructions and images, “without a line of parsing code”.

His conclusion on the hard part:

the hard part of Google and Shopee, the part where a whole Lab article ended without a working scraper, is handled on their side.

Side by Side With One of the Largest Players

A few days earlier, Pierluigi had reviewed the web scraping API of one of the largest providers in the industry, so he put the two products next to each other. That comparison is the part we are proudest of: one year in, we are being measured against companies that have been in this market for well over a decade.

He described a real difference in approach, and he described it fairly. With the larger provider you choose how each page is fetched and pay a different price for each option. With ScrapeUnblocker you send the URL, the service picks the route, and every call costs the same. Less control, a cost that is easier to predict, and which one you prefer depends on the project.

On ready-made scrapers, the catalogs overlap less than he expected: only 12 websites appear in both lists. Our 75 plugins go deep on Asian and European marketplaces (Shopee, Taobao, 1688, Zalando, Vinted, Leboncoin), travel (Booking, Airbnb, Skyscanner, Tripadvisor), real estate, jobs, review sites and SEO tools, and on Google we cover Images, Jobs, News and the Ads Transparency Center. The other catalog is stronger on US grocery chains and Latin American marketplaces. Both have a generic endpoint for any URL.

He also liked how we bill. One request is one credit, whatever the target needs, and blocks, timeouts and our own errors are free:

The billing rules are simple, and in my opinion they’re the best part of the docs.

The whole test, 35 billed requests, would have cost €0.035 at pay-as-you-go prices, well inside the free trial of 500 requests.

What He Flagged, and What We Shipped

A good review is useful because it is honest. Pierluigi listed every place where the data needed care before it goes into a dashboard. We went through the list the same week, and all of it is live now:

• Cookie banners in Markdown. On a page with a large consent manager, the “main content only” Markdown started with the cookie banner. Markdown output now drops consent managers such as Cookiebot and OneTrust before it picks the main content.
• One size’s price in parsed_data. On Shein, every size is its own offer in the page’s Schema.org data, and parsed_data returned one of them instead of the price on screen. It now takes the lowest offer, which is the price the page shows.
• Price filters. Shopee and Shein both let some promoted products through a price filter. The Shopee and Shein plugins now enforce min_price and max_price on the price they return.
• The crossed-out price on Shopee products. When a product page showed a discount without the original price, the plugin now takes it from the search listing.
• Shein review counts. Shein reports any count above 1,000 as 1001. The plugin now also returns reviewCountText (“1000+”) so you can tell the two apart.
• Google totalResults. It now reads only the result count Google shows, and the SERP guide lists the countries that are really available, Brazil included.

Each item is in our changelog.

What Comes Next

Ten calls per target is a first check, and Pierluigi says so himself. His next step is a bigger sample with concurrency, to see how the success rate holds at the limit of a plan. We are looking forward to it. For our own numbers at scale, our benchmark runs thousands of calls per target, and every scraper page shows its own measured success rate.

Thank you, Pierluigi, for the time, the care, and the honest notes.

FAQ

Who is Pierluigi Vinciguerra?

He is the founder of The Web Scraping Club, a newsletter and community about web scraping read by engineers and data teams across the industry. He writes hands-on tests of scraping tools and long technical “Lab” articles on hard targets.

How did the review come about?

We contacted Pierluigi and asked him to test the product. He chose the targets and the method, wrote his own client and published his findings, including the issues listed above.

Can I repeat the test myself?

Yes. His test scripts are linked in the article, and you can run them with a free ScrapeUnblocker account: 500 requests, no card needed.

Which endpoints did he use?

/serpApi for Google, the Shopee and Shein plugins, and /getPageSource with parsed_data=true for Shein product pages. All of them are described in the docs.

Try It on Your Own Targets

If Google, Shopee or Shein are on your list, or any other site that keeps blocking you, create a free account and send your first request in a few minutes. The docs have a quickstart for every endpoint and plugin.