Amazon Reviews Scraper

Turn a list of Amazon ASINs or product links into structured customer review data — rating, title, full review text, author, date, verified-purchase badge, helpful-vote count and purchased variant, one row per review.

Why use this actor

• No account, no login, no API key — just paste product IDs or URLs and run.
• Full review text, not just a snippet — long, multi-paragraph reviews come through complete.
• Verified-purchase badge and helpful-vote count on every review, so you can weigh reviews by trust and usefulness.
• Multi-country reviews — a US product page can surface reviews originally posted from other marketplaces (Mexico, Canada, India, etc.), and they all come through with their country tagged.
• Stable JSON output ready for spreadsheets, databases, dashboards or sentiment-analysis pipelines. Export to JSON, CSV or Excel.
• Handles blocks automatically and retries — no scrapers or browsers to babysit.

How it works

1. Enter one or more ASINs (the 10-character product code) and/or full product URLs.
2. Pick the marketplace (Amazon US, UK or Germany).
3. The actor opens each product page and reads the reviews Amazon displays directly on it — Amazon's own curated selection of top and recent reviews.
4. Each review becomes one row in your dataset.

You don't manage any of the fetching, retries, or blocking — just the list of products you care about.

Input

{
    "asins": ["B08N5WRWNW"],
    "productUrls": [],
    "domain": "amazon.com",
    "maxReviews": 0,
    "maxConcurrency": 2,
    "proxyConfiguration": {
        "useApifyProxy": true,
        "apifyProxyGroups": ["RESIDENTIAL"],
        "apifyProxyCountry": "US"
    }
}

Field	Type	Description
asins	array	Amazon product IDs, e.g. B08N5WRWNW. Add as many as you like.
productUrls	array	Full product URLs (/dp/...); the ID is detected automatically.
domain	string	Marketplace: amazon.com, amazon.co.uk or amazon.de.
maxReviews	integer	Cap the reviews collected per ASIN. 0 = collect every review Amazon shows on the page (see Notes for the typical count).
maxConcurrency	integer	How many products to fetch in parallel.
proxyConfiguration	object	Residential proxy recommended, matched to the marketplace's country (US for amazon.com).

Output

One record per review. Real sample captured from a live run on ASIN B0BDHWDR12 (Apple AirPods Pro 2nd Gen):

{
    "_source": "top-reviews-widget",
    "_scrapedAt": "2026-08-06T05:07:01Z",
    "asin": "B0BDHWDR12",
    "productUrl": "https://www.amazon.com/dp/B0BDHWDR12",
    "domain": "amazon.com",
    "reviewId": "R3PCXFKB39EVT2",
    "rating": 5,
    "title": "The king of wireless earbuds",
    "body": "These are perfect in every way possible. The sound, noise canceling, quality, battery life, everything. I originally had the airpods pro (1st gen) and the sony wf 1000 xm4...",
    "author": "Marshal H",
    "reviewDate": "2022-09-23",
    "reviewDateText": "Reviewed in the United States on September 23, 2022",
    "country": "United States",
    "verifiedPurchase": true,
    "helpfulVotes": 203,
    "variant": null
}

A negative review, from the same run, showing the country tagging on non-US reviews:

{
    "_source": "top-reviews-widget",
    "_scrapedAt": "2026-08-06T05:07:01Z",
    "asin": "B0BDHWDR12",
    "productUrl": "https://www.amazon.com/dp/B0BDHWDR12",
    "domain": "amazon.com",
    "reviewId": "R3NABKZD0H3TXG",
    "rating": 1,
    "title": "PÉSIMAS CONDICIONES",
    "body": "El producto a parte de tener un deplorable empaque, no es funcional, no cumple con las funciones de modo ambiente, cancelación de ruido, etc. He iniciado un proceso de devolución.",
    "author": "Miguel Bautista",
    "reviewDate": "2024-04-01",
    "reviewDateText": "Reviewed in Mexico on April 1, 2024",
    "country": "Mexico",
    "verifiedPurchase": true,
    "helpfulVotes": null,
    "variant": null
}

That run returned 13 full reviews for this ASIN — ratings from 1 to 5 stars, spanning the United States, Canada, Mexico and Egypt, all with complete review text.

Field	Type	Description
asin	string	The product's 10-character Amazon ID.
productUrl	string	Product page URL the review was read from.
domain	string	Marketplace the review was read from.
reviewId	string	Amazon's internal review ID.
rating	number	Star rating out of 5.
title	string	Review headline.
body	string	Full review text.
author	string	Display name of the reviewer.
reviewDate	string	Review date as YYYY-MM-DD, when it could be parsed.
reviewDateText	string	The date exactly as shown on the page (e.g. "Reviewed in the United States on September 23, 2022").
country	string	Country the review says it was posted from.
verifiedPurchase	boolean	Whether the review carries Amazon's "Verified Purchase" badge.
helpfulVotes	number | null	Number of "found this helpful" votes, when shown.
variant	string | null	The size/color/style variant the reviewer bought, when shown.

Failed lookups never fail silently — they produce a status record (_status: not_found, blocked, or no_reviews_found, with _statusDetail explaining why) so nothing disappears from your dataset.

Notes & limits

• This actor collects Amazon's "Top reviews" — the curated set of reviews (usually 5 to 15 per product) that Amazon displays directly on the product page — not a full paginated review history. We tested Amazon's dedicated review-listing pages directly and found they redirect to a sign-in page for every anonymous visitor, on the very first page, regardless of how the request is made. There is no way to read a full review archive (hundreds or thousands of reviews) without a logged-in Amazon account. If your use case needs a quick, real read on what customers are saying — sentiment, standout complaints, verified-purchase quality — this actor delivers that reliably; it is not a fit if you need every review a product has ever received.
• Use a Residential proxy set to the marketplace's country (US for amazon.com). Datacenter proxies are often shown a block page.
• Some products show zero reviews if they're new, have very few ratings, or Amazon didn't render the reviews section for that particular request — you'll get a no_reviews_found status record rather than a silent empty result.
• maxReviews rarely limits anything in practice since the page itself only carries a handful of reviews; it's there for the rare product that shows more.