Amazon Product & Reviews Scraper

Scrape Amazon product details, search-result listings, and on-page top customer reviews from server-rendered HTML. No login and no API key required. Give the actor a list of ASINs for full product details + reviews, or a list of search queries for paginated result listings. Supports 7 marketplaces (com, co.uk, de, fr, it, es, ca), and forces each marketplace's native currency/language via an i18n cookie so prices come back consistently.

What it does

• ASIN / product mode — for each ASIN, returns title, price, currency, rating, review count, brand, availability, feature bullets, hi-res images, and the top reviews embedded in the product page.
• Search mode — for each query, paginates through Amazon's result pages (up to maxItems) and returns asin, title, price, currency, rating, reviewCount, a thumbnail image (~200px), sponsored flag, and url.

Input

{
  "asins": ["B0BS1PRC4L"],
  "searchQueries": ["wireless earbuds"],
  "domain": "com",
  "maxItems": 50,
  "proxyConfiguration": { "useApifyProxy": true, "apifyProxyGroups": ["RESIDENTIAL"] }
}

Provide asins, searchQueries, or both. At least one is required — an empty input returns an uncharged BAD_INPUT diagnostic row (it does not crash the run).

Residential proxies are strongly recommended: Amazon captcha-checks datacenter IPs quickly. Use a proxy country matched to the marketplace (e.g. US residential for amazon.com).

Sample output

Product row (ASIN mode):

{
  "ok": true,
  "type": "product",
  "asin": "B0BS1PRC4L",
  "url": "https://www.amazon.com/dp/B0BS1PRC4L",
  "title": "Example Wireless Earbuds",
  "price": "$49.99",
  "currency": "USD",
  "rating": "4.5",
  "reviewCount": "1,234 ratings",
  "brand": "ExampleBrand",
  "availability": "In Stock",
  "featureBullets": ["40h battery life", "IPX5 water resistant"],
  "images": ["https://m.media-amazon.com/images/I/hi-res-1.jpg"],
  "reviews": [
    { "id": "R1", "author": "Jane", "rating": "5.0", "title": "Great", "date": "...", "verifiedPurchase": true, "helpfulVotes": "3 people found this helpful", "body": "..." }
  ],
  "reviewsParsed": 1
}

Search row (search mode):

{
  "ok": true,
  "type": "search",
  "query": "wireless earbuds",
  "asin": "B0XXXXXXXX",
  "title": "Example Wireless Earbuds",
  "price": "$49.99",
  "currency": "USD",
  "rating": "4.5",
  "reviewCount": "1,234",
  "images": ["https://m.media-amazon.com/images/I/thumb-200px.jpg"],
  "sponsored": false,
  "url": "https://www.amazon.com/dp/B0XXXXXXXX"
}

Both modes expose images as an images array so the output schema is uniform downstream. The difference is content: product mode yields hi-res image URLs, search mode yields a single ~200px thumbnail URL.

Nullable fields

These fields can be null/empty when Amazon does not render them for a given product or listing: price, rating, reviewCount, brand, availability, images (empty array), and reviews (empty array). Plan downstream pipelines to tolerate nulls.

Pricing

Pay-per-result: you are charged once per successfully parsed product/search row (ok: true). Blocked, failed, truncated, or empty results are never charged — those come back as ok: false diagnostic rows with an errorCode (BLOCKED, RATE_LIMITED, SERVER_ERROR, NETWORK, NO_RESULTS, BAD_INPUT).

Troubleshooting

• BLOCKED / robot check — Amazon served a captcha/robot-check page. Use a residential proxy (or one in the marketplace's country) and retry. Datacenter IPs get blocked quickly.
• Truncated search results — if a later page (page 2+) is blocked mid-run, the actor returns the rows it already collected and adds a diagnostic row explaining the truncation. Add/rotate a residential proxy and retry for the full set.
• NO_RESULTS — the page loaded but nothing matched your input; check the ASIN/query and the selected domain.
• BAD_INPUT — you provided neither ASINs nor search queries.