Amazon Bestsellers Scraper (7 marketplaces)

What does Amazon Bestsellers Scraper do?

It reads Amazon's public Best Sellers ranking pages and returns one row per ranked product, with the position Amazon itself printed on the page. The input is a list of category slugs and marketplaces, not a list of products: you say which shelves to read, and the run comes back with what is selling on them right now. That is what separates it from its two siblings — Amazon Product Scraper answers a question about an ASIN you already have, Amazon Search Scraper answers a keyword, and this one answers "what is at the top of this category today".

• Every marketplace you pick is crossed with every category you pick, so three marketplaces and two categories is six pages and roughly 180 rows from one run.
• Seven storefronts are available together — US, UK, Germany, France, Spain, Japan and Australia — and each row carries the marketplace it came from.
• rank is Amazon's own badge number read off the page, so it survives sorting, filtering and joining. It is not the order rows happened to arrive in.
• Every row carries an asin, which joins straight onto the Amazon Product Scraper for the full detail record.
• A category slug that does not exist on a given marketplace fails only that one target. The rest of the run still returns.

No Amazon account, API key or cookies are involved. The Actor reads the Best Sellers page a logged-out visitor sees, at about 30 products per category page and a median of 2.7 seconds per page in our own measurements.

What data does it extract?

Field	Type	Description
marketplace	string	Which storefront the row came from: us, uk, de, fr, es, jp or au
category	string	The category slug that was requested, echoed back on every row
rank	number	Bestseller position from the page's own rank badge, 1-based
asin	string	Amazon's ten-character product ID. Stable, and the right key to join or de-duplicate on
title	string	Product title, HTML entities decoded (Tom & Jerry arrives as Tom & Jerry)
price	string	The price exactly as the ranking page displayed it, symbol and grouping included
rating	number	Stars out of 5. A locale decimal comma is read as a decimal point, so 4,4 on amazon.de is 4.4 and not 44
reviewCount	number	Review total as an integer: 278974, not "278,974"
productUrl	string	Direct /dp/ link on the marketplace the row came from
sourceUrl	string	The Best Sellers page the row was read from
collectedAt	string	ISO timestamp stamped once per run

Rows that carry neither an ASIN nor a title are dropped rather than emitted half-empty, so a malformed tile in the grid costs you that tile and nothing else.

price is a string here, and deliberately so. Unlike the product and search Actors, this one does not pin the exit country per marketplace, so treat the value as the display string the page printed rather than as a currency-tagged number. When you need a numeric price with an ISO currency beside it, take the asin from these rows and run it through the Amazon Product Scraper.

Why scrape Amazon Best Sellers?

• Rank tracking. Run the same categories daily and you have a time series of who moved up, who fell off the page, and what price they were charging while it happened.
• Launch detection. A new ASIN appearing in a category's top 30 is one of the earliest public signals that a competitor's product is working.
• Cross-country assortment. The same category slug on us, de and jp shows which products travel and which are local. This is the reason the marketplace list is an array rather than a single choice.
• Sourcing and niche research. Rating and review count on every row give you a rough demand floor for a category before you commit to stock.
• Merchandising and pricing dashboards. Schedule the run, point a dashboard at the dataset, and the rank ladder for your categories refreshes itself.

How to use it

1. Open the Actor and click Try for free (or Start if you are already signed in).
2. Fill in Categories with one or more slugs. The slug is the path segment in a Best Sellers URL: amazon.com/gp/bestsellers/electronics/ is electronics, and toys-and-games and books work the same way.
3. Tick the Marketplaces you want. Each one is fetched for each category.
4. Set Maximum results if you want a hard ceiling on the dataset, and raise Concurrency if you are collecting many combinations at once.
5. Run it. When it finishes, check the SUMMARY record in the run's key-value store for the item count and for any target that failed, then export the dataset as JSON, CSV, Excel or XML, or read it through the API.

Input

Field	Type	Default	Notes
marketplaces	array of strings	["us"]	Any of us, uk, de, fr, es, jp, au. At least one is required
categories	array of strings	["electronics"]	Bestseller category slugs, 1 to 50 of them. Letters, digits and hyphens only
maxItems	integer	1000	Ceiling on dataset rows, 1 to 50,000
concurrency	integer	3	Pages fetched in parallel, 1 to 10

{
  "marketplaces": ["us", "uk", "de"],
  "categories": ["books", "electronics"],
  "maxItems": 1000,
  "concurrency": 3
}

Category slugs are marketplace-specific. books resolves on all seven storefronts; electronics is a valid slug on amazon.com but not on amazon.de, where the equivalent shelf is computers. An unknown slug fails that single target with a message saying the slug may not exist there, and every other target in the run still returns its rows.

Duplicates in either list are collapsed before any fetching happens, so the same category listed twice costs one page, not two.

Output

One row per ranked product. This is the shape rather than a captured run:

[
  {
    "marketplace": "us",
    "category": "electronics",
    "rank": 1,
    "asin": "B08JHCVHTY",
    "title": "Blink Plus Plan (Monthly Subscription)",
    "price": "$11.99",
    "rating": 4.4,
    "reviewCount": 278974,
    "productUrl": "https://www.amazon.com/dp/B08JHCVHTY",
    "sourceUrl": "https://www.amazon.com/gp/bestsellers/electronics/",
    "collectedAt": "2026-08-20T09:14:02.881Z"
  },
  {
    "marketplace": "de",
    "category": "computers",
    "rank": 2,
    "asin": "B0BDHWDR12",
    "title": "SanDisk Ultra microSDXC 128 GB",
    "price": "12,99 €",
    "rating": 4.7,
    "reviewCount": 41230,
    "productUrl": "https://www.amazon.de/dp/B0BDHWDR12",
    "sourceUrl": "https://www.amazon.de/gp/bestsellers/computers/",
    "collectedAt": "2026-08-20T09:14:02.881Z"
  }
]

Reliability

Measured across 21 marketplace-by-category combinations in a single run: 20 of 21 targets succeeded and 600 products were collected, with every schema field populated. The one failure was a category slug that does not exist on that marketplace, and it was reported as exactly that.

Ranks, prices, ratings and counts are parsed from the Best Sellers grid itself, so they are the page's own values rather than a language model's reading of them. The one place that shows is the rating line, which is localized: the parser knows the English, German, French and Spanish wordings of "out of 5 stars". Where a storefront prints it in some other form, the row still carries rank, ASIN, title and price, with rating left null instead of guessed at.

How much does it cost?

Pricing is pay-per-event: you are charged per result row written to the dataset, and nothing else. Rows are only written when they can be charged, so a run that hits your account spend limit stops writing rather than quietly handing you unbilled data. Current per-row pricing is on the Actor's Store page, which is the only place it can go stale.

Runs started from an Apify free plan stop at 250 requests and 250 results, and the run reports that it reached the limit in its status message. Any paid plan runs the full input and the maxItems you set.

That limit exists because this Actor fetches through our own scraping infrastructure, which Apify does not cover for free-plan runs. It binds on requests as well as results, so a long input list cannot spend those fetches on rows the run was never going to return.

Using it with the Apify API

import { ApifyClient } from 'apify-client';

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

const run = await client.actor('usestring/amazon-bestsellers').call({
  marketplaces: ['us', 'de'],
  categories: ['books'],
  maxItems: 500
});

const { items } = await client.dataset(run.defaultDatasetId).listItems();
console.log(items.slice(0, 3));

The same run is reachable from the Python client, from any HTTP client against the Apify REST API, and from a schedule if you want the ranking snapshot taken for you every morning.

Limitations

Ranking pages only. There is no product description, image set, variant matrix, seller list, review text or Q&A here — the Amazon Product Scraper covers the detail page, and the Amazon Search Scraper covers keyword results. The Actor reads the first ranking page of a category and does not paginate into the rest of the top 100. price is a display string rather than a normalized number. Marketplaces beyond the seven listed above, amazon.in included, are not supported.

Related actors

• Amazon Product Scraper — feed it the ASINs from these rows for numeric prices, currency, availability and brand.
• Amazon Search Scraper — keyword results with ranks and a sponsored flag, for the demand you cannot see from a category page.
• Walmart Product Scraper — the same product research on Walmart's catalogue, including marketplace sellers.
• Home Depot Product Scraper — price and stock for home-improvement SKUs, resolvable per store.
• Etsy Listings Scraper — handmade and vintage listings, useful when a category's competition sits outside the big-box marketplaces.

FAQ

How many products does one category return? About 30, which is one Amazon Best Sellers page. The Actor does not follow the pagination into the rest of the top 100, so a category is one fetch.

How do I find the right category slug? Open the Best Sellers page you care about on the marketplace you care about and copy the segment after /gp/bestsellers/. Slugs are per-storefront, so check them on each one rather than assuming the US slug carries over.

Which marketplaces can it collect? Seven: US, UK, Germany, France, Spain, Japan and Australia, all in one run. amazon.in is deliberately not offered.

Is the price in local currency? It is whatever the ranking page printed, which is normally the storefront's own currency. This Actor does not pin an exit country per marketplace, so if you need a guaranteed local numeric price take the ASIN over to the Amazon Product Scraper, which does.

How fresh is the data? It is the page as it stood when the run fetched it, stamped in collectedAt on every row. Amazon recalculates Best Sellers hourly, so schedule the run at whatever interval your analysis actually needs.

What happens when a target fails? The failure is recorded under failures in the run's SUMMARY record with the URL and the reason, and the other targets still write their rows. A run in which every single target failed exits as a failed run rather than as an empty success.

Do I need proxies or an anti-bot setup? No. Fetching goes through our own infrastructure and there is nothing to configure — no proxy groups, no session pool, no retry policy to tune.

Feedback

Something wrong, or a field you need that is not here? Open an issue from the Actor's Store page. Bug reports that name the exact marketplace and category slug get fixed fastest.