Amazon Bestsellers Scraper

Point it at any Amazon Best Sellers, New Releases or Movers & Shakers list and get the ranking back as a table: position, ASIN, title, price, star rating, review count, image, product link and the category path. Paste the URLs, or just type department names and pick a marketplace.

No account, no cookies, no login, no browser.

• Works from the URL you copied out of the address bar, in either link shape Amazon uses.
• Or skip URLs entirely. Type "Home & Kitchen" and it resolves to the right list on its own.
• 22 marketplaces, from amazon.com to amazon.co.jp. You can mix countries in one run.
• Prices come back in the marketplace's own currency, not converted for wherever the request left from.
• Empty input returns one labelled sample row, free.

Price

$0.26 per 1,000 products. Plus $0.0007 per run as the platform start fee, billed per gigabyte of run memory, so exactly that on the default 1 GB.

No volume tiers, no minimum spend, no subscription.

Products	Total
100	$0.0267
1,000	$0.2607
10,000	$2.6007
100,000	$26.0007

What gets charged

• One product-scraped event per product row written to the dataset. Nothing else is metered per row.
• Free: the sample row an empty run returns, and every diagnostic row. They all carry "charged": false.
• A category Amazon returns empty (which Movers & Shakers usually is) costs nothing beyond the start fee.
• A category name that doesn't exist on the marketplace you chose gets a free diagnostic row saying so.
• The same product appearing twice inside one category list is dropped before it's charged.
• A run that finds nothing costs the start fee and nothing else.
• Rows never leave the dataset without a charge and are never charged without a row. The billed event is a named one, so nothing is quietly attached to apify-default-dataset-item.

Input

{
  "categoryUrls": [
    "https://www.amazon.com/gp/bestsellers/electronics/",
    "https://www.amazon.co.uk/Best-Sellers-Books/zgbs/books/"
  ],
  "categories": [
    "Home & Kitchen",
    "Toys & Games"
  ],
  "listType": "bestsellers",
  "domain": "com",
  "maxItemsPerCategory": 60
}

Field	What it does
categoryUrls	Amazon list URLs copied straight from the site. Both shapes work: /gp/bestsellers/electronics/ and /Best-Sellers-Electronics-Headphones/zgbs/electronics/172541/. So do /gp/new-releases/... and /gp/movers-and-shakers/.... Each URL carries its own marketplace and list type, so you can mix countries and list types freely. Up to 50 per run.
categories	Department names written the way Amazon writes them in that marketplace's own language ("Home & Kitchen", "Toys & Games", "Bücher"), or the slug out of a URL ("electronics", "toys-and-games", "books/2"). These use the marketplace and list type set below. Up to 50 per run.
listType	bestsellers, new-releases or movers-and-shakers. Applies only to the names in categories. A URL already says which list it is.
domain	Which marketplace the names in categories belong to. Ignored for categoryUrls.
maxItemsPerCategory	How many ranked products per category, counting from rank 1. 60 is both the default and the ceiling, because that's everything Amazon renders. Lower it while testing, since you pay per product row.
proxyUrls	Leave empty. Only for callers who want traffic to leave through proxy servers they already pay for, as http://user:pass@host:port.

Run it with empty input and you get one labelled sample row, free.

Output

One row per product. This is a real row from a real run:

{
  "ok": true,
  "charged": true,
  "recordType": "product",
  "categoryName": "Electronics",
  "categoryPath": "Electronics",
  "listType": "bestsellers",
  "marketplace": "amazon.com",
  "inputUrl": "https://www.amazon.com/gp/bestsellers/electronics/",
  "rank": 2,
  "asin": "B0DCH8VDXF",
  "title": "Apple EarPods Headphones with USB-C Plug, Wired Ear Buds with Built-in Remote to Control Music, Phone Calls, and Volume",
  "brand": null,
  "price": "$16.91",
  "priceValue": 16.91,
  "currency": "USD",
  "rating": 4.5,
  "reviewCount": 16728,
  "image": "https://images-na.ssl-images-amazon.com/images/I/513OSdW4elL._AC_UL900_SR900,600_.jpg",
  "productUrl": "https://www.amazon.com/dp/B0DCH8VDXF",
  "previousRank": null,
  "rankChangePercent": null,
  "page": 1,
  "scrapedAt": "2026-08-16T13:29:52.126Z"
}

Field notes

• rank is the position Amazon prints on the badge, not a number counted here. Page 1 carries ranks 1-30 and page 2 carries 51-80, which is exactly what Amazon renders.
• asin is Amazon's product id. Stable, and the right key for diffing one run against the next.
• brand is the byline Amazon prints on the ranking card: author for a book, artist for music, studio for a film. Cards for hard goods carry no byline at all, so this is null for most of Electronics, Home & Kitchen and so on.
• price is exactly the text on the card, symbol and all. For books showing several formats it's the "offers from" price, which is the lowest of them.
• priceValue is the same figure as a number, with the marketplace's separators resolved, so 1 234,56 zł becomes 1234.56.
• currency is read from the price text when it names one, otherwise from the marketplace. It always describes the price string in the same row.
• rating is stars out of five at read time, or null when nothing has been rated yet.
• reviewCount is the ratings count next to the stars, as a number. Null when there are none.
• image is the largest size Amazon offers for that card, pulled from the card's own image set rather than the small default.
• categoryPath is the ancestor chain down to the list you asked for, like Electronics > Headphones. For a top-level department it's just the category name.
• previousRank is the rank 24 hours earlier and rankChangePercent the size of the jump. Both are Movers & Shakers fields and are null on Best Sellers and New Releases rows, which publish no rank movement.
• page is 1 for ranks 1-30 and 2 for ranks 51-80. Useful when you want only the top of a list.

Real rows carry "charged": true. Sample rows carry "_sample": true, diagnostic rows carry "_diagnostic": true and an errorCode. Neither is billed.

How it works

The ranking is read straight out of the logged-out list page, which Amazon renders on its own servers. One request per 30 products, no page rendering, no headless browser, no login.

Requests leave through a rotating pool of addresses, and a refused address is answered by moving to another one rather than by waiting. The marketplace's own currency and language preference goes out with every request, so a list read from another country still comes back priced the way a local shopper sees it.

Department names are matched against the marketplace's own Best Sellers menu rather than a hard-coded list. That's why "Home & Kitchen" finds the right list even though its address says something else entirely, and why the names work in German, Japanese or Spanish.

Each category is paged until maxItemsPerCategory is reached, and a product id already seen in that category is skipped.

What people use it for

• Product research. Pull the top 60 of a category every morning and watch which ASINs enter and leave.
• Tracking your own product, or somebody else's, by ASIN across a category and its subcategories over time.
• Sourcing decisions. The ranking, the price band and the review counts in one table.
• Spotting new entrants early with the New Releases list, which is the same run with one field changed.
• Price-band research. priceValue and currency are already typed, so the median price of a category is one spreadsheet formula.
• Comparing the same category across marketplaces. The same URL pattern works on all 22, and each row says which marketplace it came from.

Reading the output

• Real rows carry "charged": true and "recordType": "product". One billed event each.
• The sample row carries "_sample": true and "charged": false. There's exactly one, only when the input named no category at all.
• Diagnostic rows carry "_diagnostic": true, "charged": false and an errorCode: NO_RESULTS when Amazon returned the list page with no products on it, BAD_INPUT when a URL isn't an Amazon list link or a name matches no department, NOT_FOUND when the category doesn't exist, BLOCKED when Amazon refused every address tried, TIME_BUDGET when time ran out. Each carries a plain-English error and the inputUrl it belongs to.

Filter on charged == true and you have exactly the rows you paid for.

Limits

• Amazon renders 30 products per list page and links exactly two pages, so a list tops out at 60 products and the ranks you get are 1-30 and 51-80. Ranks 31-50 and 81-100 aren't rendered on Amazon's own pages, so nothing can read them from this surface.
• Movers & Shakers is supported, but Amazon has been answering most Movers & Shakers categories with an empty list. When that happens you get a free NO_RESULTS row and no charge. The run is fine. The list genuinely has nothing on it.
• brand is null for most hard goods. Amazon doesn't print a byline on those ranking cards. It isn't a parsing gap, and filling it would mean opening every product page.
• Only what the ranking card shows comes back. Bullet points, descriptions, variations, seller, stock and shipping live on the product page and are out of scope, because reading them would mean one request per product instead of one per thirty.
• Prices are a snapshot at read time and move constantly. A book with several formats shows the lowest "offers from" price, which isn't always the format at that rank.
• Best Sellers is recalculated hourly and New Releases changes daily, so two runs an hour apart legitimately disagree.
• A category name is matched against that marketplace's Best Sellers menu, which lists departments. A deeper subcategory ("Headphones", "Espresso Machines") isn't in that menu. Use its URL, or the slug/node form, for those.
• Kindle, Audible, Apps and Digital Music lists are readable, but their cards carry less than a physical product's. Some have no price, and formats vary.
• Hard ceilings: 50 categories per run and 60 products per category.
• Amazon sometimes answers with a short interstitial page instead of the list. The run moves to another address. If every attempt is refused you get an uncharged BLOCKED row rather than a partial list presented as complete.

Questions

Do I need a category URL, or can I just say what I want?

Either. categoryUrls takes links copied from Amazon. categories takes department names as Amazon writes them, in that marketplace's own language, and resolves them against Amazon's own menu. Both work in the same run.

Why do the ranks jump from 30 to 51?

Because Amazon's own Best Sellers pages do. Page one renders 30 of its 50-product block and page two starts again at 51. Open the list in a browser and you'll see the same gap. Everything here is a product Amazon actually rendered, with the rank Amazon printed on it.

Can I get more than 60 products from one category?

Not from this surface. 60 is everything Amazon renders. To go deeper, scrape the subcategories. Each has its own list of 60, and categoryPath tells you which is which.

What if a category returns nothing?

One uncharged diagnostic row for that category explaining why, and the run carries on to the others.

Do I need a proxy?

No. The run brings its own egress and its cost is already inside the price. proxyUrls exists only for callers who want traffic to leave through servers they already own.

Will the run fail if Amazon blocks it?

No. It moves to a different address and tries again. If everything is refused you get an uncharged diagnostic row and the run still finishes as succeeded. A failed run would still bill the start fee, which would mean paying to be told something went wrong.

Can I run it on a schedule?

Yes. Nothing is held between runs. Diff on asin to see what entered or left the list, and on rank to see what moved.