Amazon Seller & Shop Scraper

Scrape Amazon third-party seller intelligence — the seller profile page (business name, address, phone, star rating, feedback windows, ratings histogram) and the seller's storefront product catalog (/s?me=), in one actor. Built for lead generation, competitor research, and seller due-diligence use cases. No login, no cookies required.

What this actor does

• Three modes: PROFILE (seller profile only), PRODUCTS (storefront catalog only), BOTH (default)
• Two input paths: supply seller IDs directly, or supply product URLs and the actor auto-derives the seller ID from each product's buy-box "sold by" link
• Seller feedback windows: 30-day, 90-day, 12-month, and lifetime rating + ratings-count, whichever windows Amazon actually shows for that seller
• Feedback histogram: percentage breakdown across 5/4/3/2/1 stars
• Public business data: legal business name, business address, and customer-service phone number — data Amazon itself displays publicly on the seller profile page (toggle off with includeBusinessDetails)
• Storefront catalog: reuses the standard Amazon search-result card schema (title, price, rating, image, badges) for every product in the seller's shop
• 23 Amazon marketplaces supported (all browser-verified)
• Empty fields are omitted — no null placeholders, no fabricated data

Output: seller profile record (recordType: "sellerProfile")

• sellerId, sellerName, profileUrl, storefrontUrl
• starRating — overall star rating shown in the page header
• positivePercent12m, ratingsCount12m — 12-month positive-feedback percent and count (when shown)
• positivePercentWindow, positivePercent, positivePercentRatingsCount — the exact window Amazon's own page header currently displays (usually 12 months, but not guaranteed)
• feedbackWindows[] — {window, starRating, ratingsCount} for each of 30 days / 90 days / 12 months / lifetime that Amazon actually shows data for
• feedbackHistogram — {"5": {percent}, "4": {percent}, ...}
• aboutText — the seller's "About" description
• phone — customer-service phone number (public, includeBusinessDetails=true)
• businessName, businessAddress — public legal business identity (includeBusinessDetails=true). For the common 5-line Western format, businessAddress is {street, city, state, zip, country}; for addresses that don't fit that exact 5-line shape (common among non-Western/international sellers), businessAddress is instead {addressLines: [...]} — the real, ordered address lines exactly as Amazon renders them, never force-fit into the wrong field
• returnRefundPolicyText, shippingPolicyText, privacyPolicyText — the seller's own return/refund, shipping, and privacy-notice text, when Amazon's page renders one
• scrapedAt

Output: seller product record (recordType: "sellerProduct")

• asin, sellerId, title, url
• price {value, currency, display}, listPrice
• stars, reviewsCount
• thumbnailImage
• prime, sponsored, resultType (ORGANIC/SPONSORED), badge
• boughtInPastMonth
• position {page, positionOnPage, globalPosition}
• scrapedAt

Output: storefront facets record (recordType: "storefrontFacets")

One record per seller (mode PRODUCTS/BOTH), built from the storefront's own refinement rail and sort dropdown — lets you discover what departments/filters and sort orders exist for a seller's catalog before deciding what to request.

• sellerId, storefrontUrl
• facets[] — {groupName, label, url, selected, refinementParameter?} from the storefront's #s-refinements rail (Department, Delivery Day, Free Shipping, etc.)
• sortOptions[] — {value, label, selected} from the storefront's own "Sort by" dropdown — value is the exact string the sortBy input accepts
• pagesRequested, pagesFetched — how many storefront pages were actually fetched vs. maxProductPages requested. Usually equal; pagesFetched can be lower either because the storefront's own last page was reached naturally (fewer sellers have hundreds of pages of stock) or, rarely, because a persistent anti-bot block stopped the fetch partway through (in that case a pagesStoppedReason: "BLOCKED" field is also present, so a short run is never silently indistinguishable from a genuinely small storefront)
• scrapedAt

Output: error/terminal record (recordType: "sellerError")

Emitted instead of a normal record when a seller/page could not be resolved (not found, blocked, CAPTCHA, etc.) — the actor never silently returns zero items for a requested seller without explaining why.

• sellerId, pageClassification (e.g. NOT_FOUND, CAPTCHA, DOG_PAGE_503), scrapedAt

Input

Field	Type	Default	Description
mode	string	BOTH	PROFILE / PRODUCTS / BOTH
sellerIds	array	–	Seller/merchant IDs to look up directly
productUrls	array	–	Product URLs to derive seller IDs from (buy-box seller link)
marketplace	string	US	Amazon marketplace/domain
maxProductPages	int	3	Max storefront pages per seller (mode=PRODUCTS/BOTH), ~16 products/page
includeBusinessDetails	bool	true	Include public business name/address/phone
sortBy	string	(default/Featured)	Storefront sort order (mode=PRODUCTS/BOTH) -- matches Amazon's own "Sort by" dropdown values
useResidentialProxy	bool	false	Force residential proxy from the start
proxyCountry	string	AUTO	Proxy geo-targeting country

At least one of sellerIds or productUrls is required.

Example: look up two sellers directly

{
  "mode": "BOTH",
  "sellerIds": ["A294P4X9EWVXLJ", "A35BKAA6VXTYCT"],
  "maxProductPages": 2
}

Example: derive the seller from a product page

{
  "mode": "PROFILE",
  "productUrls": ["https://www.amazon.com/dp/B0BTYCRJSS"]
}

Example: storefront catalog only, UK marketplace

{
  "mode": "PRODUCTS",
  "sellerIds": ["A294P4X9EWVXLJ"],
  "marketplace": "GB",
  "maxProductPages": 5
}

Use cases

• Lead generation — build a list of active third-party sellers with public business name/address/phone for outreach
• Competitor research — monitor a competing seller's full storefront catalog and pricing
• Seller due-diligence — check a seller's rating history across multiple feedback windows before doing business with them
• Marketplace research — compare seller reputations (histogram, positive %) across a category
• Supply-chain verification — confirm a seller's legal business identity before sourcing

FAQ

Is this affiliated with Amazon? No. This is an independent third-party actor that scrapes Amazon's own public seller profile and storefront pages — no API keys, cookies, or login required.

Is the business name/address/phone customer personal data? No. This is the seller's own public business identity that Amazon displays on every seller's public profile page (/sp?seller=...), distinct from any Amazon customer's personal/GDPR-sensitive data. You can disable it with includeBusinessDetails: false.

Why do some sellers have fewer feedback windows than others? Amazon only shows a feedback window (30 days / 90 days / 12 months / lifetime) when the seller has ratings data for that window. A newer seller may only show 90-day and lifetime, for example — the actor reports exactly what Amazon displays, never a fabricated shape.

Does this include individual written feedback comments (per-buyer text/author/date)? No. Amazon's public seller profile page (/sp?seller=...) only ever renders the aggregate feedback data this actor extracts — star rating, positive-percent, and the 5/4/3/2/1-star histogram per time window. It does not render individual buyer feedback comments/text/author/date anywhere on that page (verified live), so the actor doesn't fabricate a per-comment feedback list.

Why is a seller record sometimes missing instead of empty? If a seller ID can't be resolved (not found, blocked, CAPTCHA), the actor emits an explicit error/terminal record (recordType: "sellerError") with a pageClassification explaining why, rather than silently returning nothing.

How many products does the storefront catalog return? Up to maxProductPages × ~16 products per seller — exactly what Amazon's own storefront pagination currently exposes for that seller, not a fixed/guaranteed count.

Does this work on every Amazon marketplace? Yes — all 23 supported marketplaces work end-to-end, including ones that show an interactive anti-bot challenge on the seller/storefront pages (it resolves automatically after the actor's normal page-load wait). Non-US marketplaces can occasionally see proxy-routing timeouts; if a run fails on a non-US marketplace, retry it and/or enable useResidentialProxy: true.

Why does profileUrl/storefrontUrl return an error in a plain curl/script check? Amazon's edge WAF blocks non-browser requests (no JS execution, no real browser TLS fingerprint) to /sp?seller=... and /s?me=... with a 503, even though the exact same URL loads fine in a real browser (or via this actor's own Playwright session — verified live). productUrl (/dp/ASIN) and media/thumbnail URLs are not behind this same block and return 200 from a plain client.

Is sortBy guaranteed across every storefront page? Amazon's own storefront reliably honors a non-default sortBy on page 1 of the catalog (verified live: price-asc-rank/price-desc-rank correctly order page 1's results). For maxProductPages > 1, the actor carries the sort session forward using Amazon's own "next page" link and query-id where available, but Amazon's storefront backend does not guarantee the same strict order past page 1 for every sort value — treat page 1 as the authoritative sorted sample if you need a strict global order.

More Amazon Actors

Part of a full Amazon scraper suite — also check out:

• Amazon Category Scraper
• Amazon Product Scraper
• Amazon Best Sellers & Rankings Scraper
• Amazon Reviews Scraper
• Amazon Reviews Scraper Pro
• Amazon Offers Scraper (All Offers Display)
• Amazon Deals Scraper
• Amazon Brand Store Scraper
• Amazon Creator Shop Scraper
• Amazon Wishlist & Registry Scraper
• Amazon Keyword Suggestions Scraper