What does it cost to scrape Amazon?

Published Amazon meters run from $0.0699 per 1,000 requests at full utilization of Scrape.do Advanced to $2.94 per 1,000 regular Amazon Product API requests at full utilization of ScrapingBee Freelance among the self-serve credit plans reviewed. That range is not the bill for buying only 1,000 pages: those two plans cost $699 and $49 respectively. Bright Data instead publishes $1.50 per 1,000 delivered Amazon records with a recurring 5,000-record free tier, while Oxylabs starts its paid Amazon result ladder at $0.50 per 1,000 with a $49 monthly minimum. (Scrape.do, ScrapingBee, Bright Data, and Oxylabs, retrieved Jul. 26, 2026.)

Six of the eight vendors compared publish an explicit Amazon meter: Bright Data, Oxylabs, ScraperAPI, ScrapingBee, Scrape.do, and ScrapingDog. Decodo publishes the premium-pool rate inherited by its Amazon templates but no separate Amazon multiplier. Crawlbase publishes Amazon scrapers and complexity-based billing, but its static pricing page does not reveal the numeric Amazon tier. (Bright Data, Oxylabs, ScraperAPI, ScrapingBee, Scrape.do, ScrapingDog, Decodo, and Crawlbase, retrieved Jul. 26, 2026.)

The numbers, with the purchase floor left in

“Normalized paid rate” below is the plan price divided by the maximum Amazon requests or records that its published allowance covers. It assumes the allowance is fully used. “Minimum paid purchase” is the smallest listed paid plan, not a claim that a buyer sending only 1,000 inputs consumes its whole allowance. Free tiers and trials are stated separately. Every calculation uses monthly list prices, excludes taxes, and does not assume an unlisted overage. (ScraperAPI, ScrapingBee, Scrape.do, and ScrapingDog, retrieved Jul. 26, 2026.)

The numbers, with the purchase floor left in table

Bright Data’s billing unit is a record, not a page. Its product scraper can return a product record, while separate scrapers cover search results, reviews, sellers, and best-seller inputs. A job that submits 1,000 search pages can therefore deliver more than 1,000 records, so $1.50 per 1,000 records must not be relabeled as $1.50 per 1,000 pages. (Bright Data Amazon Scraper, retrieved Jul. 26, 2026.)

Oxylabs also bills a successful result and publishes plan-specific Amazon rates. The $0.25 Custom+ figure is a starting unit rate without a published monthly commitment, so it is a quote-level rate, not a $0.25 self-serve purchase. (Oxylabs pricing, retrieved Jul. 26, 2026.)

Worked budget: one request beyond Scrape.do Hobby

Assume a job makes exactly 250,001 successful Amazon plugin requests in one month. Scrape.do records one credit per successful Amazon request. Its Hobby plan costs $29 and includes 250,000 credits, while Pro costs $99 and includes 1,250,000 credits. The dataset records no numeric overage price for either plan, so the example selects Pro rather than inventing a top-up price. (Scrape.do plans and Amazon request cost, retrieved Jul. 26, 2026.)

Worked budget: one request beyond Scrape.do Hobby table

The $70 difference and 999,999-credit remainder are calculated totals, not vendor-listed prices or allowances. The listed inputs are the $29 and $99 monthly prices, the 250,000-credit and 1,250,000-credit allowances, and the one-credit Amazon meter. An extra-credit purchase could change the choice, but Scrape.do does not publish its numeric price in the reviewed record. (Scrape.do pricing and request costs, retrieved Jul. 26, 2026.)

What a dedicated Amazon endpoint changes

The dedicated products do more than relay Amazon HTML:

• ScrapingBee publishes product, search, and pricing endpoints that return formatted JSON. Product and pricing light calls cost 5 credits and regular browser-rendered calls cost 15. The same documentation contains conflicting search-light text, showing both 5 and 10 credits in different passages, so the table does not claim a search-light cost. (ScrapingBee Amazon API, retrieved Jul. 26, 2026.)
• Scrape.do publishes product-detail, offer-listing, and search endpoints that return parsed JSON, plus a raw-HTML option. Each successful Amazon plugin call costs 1 credit. (Scrape.do Amazon API, retrieved Jul. 26, 2026.)
• Oxylabs exposes separate Amazon product, search, pricing, seller, and best-seller targets rather than one undifferentiated URL fetch. Its target-specific rate replaces the “other sites” rate. (Oxylabs targets and pricing, retrieved Jul. 26, 2026.)
• Decodo lists Amazon best-seller, pricing, product, search, seller, and arbitrary-page templates. Target templates use its Premium Proxy Pool, which is why the premium result rate, rather than a separately named Amazon rate, appears in the table. (Decodo template catalog and pricing, retrieved Jul. 26, 2026.)
• Crawlbase publishes separate scrapers for product details, search results, offers, reviews, best sellers, and new releases. Those scrapers return structured JSON through the Crawling API, but the public static calculator does not expose a numeric Amazon price. (Crawlbase scraper catalog and pricing, retrieved Jul. 26, 2026.)
• Bright Data meters successfully delivered structured records and includes proxy management, browser rendering, CAPTCHA handling, parsing, batch collection, and job delivery in the listed Amazon endpoint. (Bright Data Amazon Scraper, retrieved Jul. 26, 2026.)
• ScrapingDog publishes separate product, search, review, autocomplete, and offers APIs. Its public credit table prices Product and Search at 1 credit each and Offers at 5, but does not list separate figures for Reviews or Autocomplete. (ScrapingDog documentation, retrieved Jul. 26, 2026.)

ScraperAPI is the exception among the vendors with an explicit Amazon meter: its reviewed documentation applies a 5-credit Amazon domain cost to the general request API rather than publishing a separate Amazon endpoint. (ScraperAPI credit costs, retrieved Jul. 26, 2026.)

That output difference matters to cost. A product endpoint can remove a separate parsing job, but a record meter can expand with the number of products returned. A raw-page endpoint keeps the bill attached to requests, but the buyer still owns the page-to-fields transformation. The published units above allow those two products to be compared without pretending they are identical. (Bright Data, ScrapingBee, and Scrape.do, retrieved Jul. 26, 2026.)

Where the protection cost appears

The clearest evidence is in the vendors’ meters. ScraperAPI prices an Amazon request at 5 credits against a 1-credit ordinary request. ScrapingBee’s Amazon Product API light and rendered routes cost 5 and 15 credits. By contrast, Scrape.do prices every successful Amazon plugin request at 1 credit, while ScrapingDog prices Product and Search at 1 credit and Offers at 5. Those are pricing facts, not measurements of speed or reliability. (ScraperAPI credit costs, ScrapingBee Amazon API, Scrape.do Amazon API, and ScrapingDog documentation, retrieved Jul. 26, 2026.)

Bright Data and Scrape.do both describe CAPTCHA handling, proxy management or sessions, location-aware collection, and browser-capable routes as part of their Amazon products. ScrapingBee explicitly ties the 10-credit difference between its light and regular Amazon calls to browser rendering. These mechanisms explain why a generic 1-credit HTML rate is not a defensible Amazon budget. They do not establish that every Amazon URL presents every protection on every request. (Bright Data Amazon Scraper, Scrape.do Amazon API, and ScrapingBee Amazon API, retrieved Jul. 26, 2026.)

The official route is Creators API, not PA-API 5

Amazon states that Product Advertising API 5.0 is deprecated and is being replaced by Creators API. The deprecation notice says PA-API 5 calls can receive HTTP 403 AccessDeniedException, so current buying advice that presents PA-API 5 as the supported option is out of date. (Amazon PA-API 5 deprecation notice, retrieved Jul. 26, 2026.)

Creators API returns Amazon catalog data through SearchItems, GetItems, GetVariations, and GetBrowseNodes. It is intended for publishers, influencers, and affiliate partners building shopping experiences. Access requires enrollment in the relevant Amazon Associates program, at least 10 qualifying sales in the previous 30 days, registration, and new API credentials. Amazon publishes no separate per-call dollar price on the reviewed introduction or rate page. (Creators API introduction and API rates, retrieved Jul. 26, 2026.)

Initial Creators API credentials receive up to 1 transaction per second and 8,640 transactions per day for the first 30-day period; later allocation is tied to shipped revenue. One GetItems call containing 10 ASINs still counts as one transaction. (Creators API rates, retrieved Jul. 26, 2026.)

Creators API answers the same question as a scraper only when the job fits its four catalog operations. Amazon documents no Creators API operation for arbitrary rendered HTML, review pages, seller pages, or the exact retail search interface, so it is not a general page-retrieval substitute. The migration also changes authentication from AWS Signature Version 4 to OAuth 2.0 and requires new credentials. (Creators API introduction and migration guide, retrieved Jul. 26, 2026.)

Amazon’s published position on automated access

As of Jul. 26, 2026, Amazon’s Conditions of Use says its license does not include collection and use of product listings, descriptions, or prices, or use of data-mining tools, robots, or similar data-gathering and extraction tools; this sentence reports the document and is not legal advice. (Amazon Conditions of Use, retrieved Jul. 26, 2026.)

The cost decision is therefore output-specific. Use Creators API when its eligible affiliate catalog operations answer the job. For a page or entity outside that response contract, compare the scraper’s real billing unit, target meter, and purchase floor, and review the applicable terms before collection. (Creators API introduction, Bright Data Amazon Scraper, and Amazon Conditions of Use, retrieved Jul. 26, 2026.)