Amazon Reviews Scraper — ratings, verified purchases & keyword summaries

Turn supported Amazon ASINs and product URLs into analysis-ready review rows. Export available public review text, normalized ratings, verified-purchase status, helpful votes, variants, stable IDs, and timestamps to JSON, CSV, Excel, or the Apify API.

The Actor charges $0.003 per successful Dataset review ($3 per 1,000) plus a $0.00005 run-start event. maxResults is a hard run-wide cap, so you can bound review charges before starting.

Why use it?

• Accept exact ASINs or supported Amazon product/review URLs.
• Process up to 25 products per run across 12 selectable marketplaces.
• Deduplicate products and reviews before paid Dataset delivery.
• Normalize ratingValue, verifiedPurchaseValue, and helpfulCountValue while preserving displayed source values.
• Add a stable reviewId, canonical product URL, source mode, and UTC collectedAt timestamp.
• Store rejected inputs, unavailable products, run totals, and product summaries in the free Key-Value Store.
• Reject spoofed Amazon hosts and arbitrary strings that merely contain an ASIN-like token.

Availability varies by product, marketplace, Amazon page state, region, and time. Page 1 uses the public product page; deeper review pages are best effort and may be unavailable.

Quick start

{
  "asins": ["B09B8V1LZ3"],
  "domain": "amazon.com",
  "maxPages": 1,
  "maxResults": 25
}

URLs select their own marketplace:

{
  "asins": [
    "https://www.amazon.co.uk/dp/B08L5WHFHY",
    "https://www.amazon.de/product-reviews/B07PGL2ZSL"
  ],
  "maxResults": 50,
  "sortBy": "recent",
  "filterByStar": "all_stars"
}

Supported marketplace selections are Amazon US, UK, Germany, France, Italy, Spain, Canada, Australia, Japan, India, Mexico, and Brazil. Support means the input and URL form are recognized; it does not guarantee that Amazon exposes reviews for every request.

Input reference

Field	Type	Default	Description
asins	string[]	[]	Up to 25 exact ASINs or supported Amazon product/review URLs.
domain	string	amazon.com	Marketplace used for plain ASINs.
maxResults	integer	25	Hard cap of 1–100 paid Dataset reviews across the run.
maxPages	integer	1	Best-effort pages per product, from 1–5.
sortBy	string	recent	Requested recent or helpful order.
filterByStar	string	all_stars	Requested star filter on review-page requests.
extract*	boolean	true	Toggle displayed reviewer, rating, title, date, text, verified, helpful, country, and variant fields.
maxConcurrency	integer	2	Products processed in parallel, from 1–5.
timeout	integer	120	Per-page timeout, from 45–300 seconds.

Dataset output

Only substantive, deduplicated review records enter the paid Dataset. A typical row looks like this:

{
  "success": true,
  "asin": "B09B8V1LZ3",
  "domain": "amazon.com",
  "productUrl": "https://www.amazon.com/dp/B09B8V1LZ3",
  "reviewId": "B09B8V1LZ3-4a58f3d7a2c219cc1e891b6a",
  "sourceMode": "product_page",
  "reviewerName": "Amazon Customer",
  "rating": "5.0 out of 5 stars",
  "ratingValue": 5.0,
  "title": "Clear sound, simple setup",
  "date": "Reviewed in the United States on August 20, 2026",
  "text": "The displayed public review text...",
  "verifiedPurchase": "Verified Purchase",
  "verifiedPurchaseValue": true,
  "helpfulCount": "18 people found this helpful",
  "helpfulCountValue": 18,
  "productVariant": "Color: Charcoal",
  "collectedAt": "2026-08-27T12:00:00Z"
}

Fields can be blank or null when Amazon does not display them. The normalized fields make filtering and aggregation easier without hiding the original displayed values.

Free Key-Value Store reports

• SUMMARY — requested, valid, invalid, duplicate, failed, delivered, and cap counts.
• PRODUCT_SUMMARIES — rating histogram, average rating, verified-status coverage, and English keyword signals for the reviews actually delivered.
• FAILED_INPUTS — invalid, unsupported, or excess inputs; these do not enter the paid Dataset.
• FAILED_PRODUCTS — valid products with no substantive results and bounded failure diagnostics.

sentimentScore, topPositiveWords, and topNegativeWords are a transparent English keyword heuristic. They are directional aids, not model-based sentiment, customer intent, product quality, or a factual verdict. Small samples and non-English reviews can be misleading.

Pricing and billing safety

• Actor start: $0.00005 per run.
• Successful Dataset review: $0.003.
• Empty, invalid, duplicate, failed, and diagnostic records are not added as paid Dataset items.
• maxResults limits paid review rows across all products in the run.
• Stable IDs and deduplication help reconcile repeated runs, but Amazon can change displayed content.

API example

curl -X POST "https://api.apify.com/v2/acts/KXlDi1LT7eCGlCmRs/runs?token=YOUR_APIFY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"asins":["B09B8V1LZ3"],"maxPages":1,"maxResults":10}'

You can also install n8n-nodes-apivault-amazon-reviews from n8n Community Nodes and run the Actor with your Apify token.

Responsible use

Reviews can contain personal data and copyrighted text. Collect only what you need, respect applicable laws and platform terms, avoid profiling or harassment, and do not republish large bodies of review text without the necessary rights. Verify important conclusions against the source. Do not use keyword summaries alone for high-impact decisions about people.

Amazon is a trademark of Amazon.com, Inc. This independent Actor is not affiliated with or endorsed by Amazon.

Reliability notes

• Public page structure and availability can change without notice.
• Requested sorting and star filters depend on Amazon exposing the corresponding review page.
• Page 1 is usually the best starting point; deeper pages may return no additional rows.
• A successful run can legitimately contain zero paid Dataset rows. Check SUMMARY and FAILED_PRODUCTS for the reason.