Amazon Reviews Scraper: Ratings, Text & Verified Badge

The Mine Works: 544 users across 100 public Actors, >99% run success. Pay only for results delivered. Browse all Actors.

Overview

Amazon Reviews Scraper turns any ASIN into clean, structured customer-review JSON: star rating, review title and body, verified-purchase badge, helpful votes, review images, author name, product variant, and date. No Amazon seller account, no SP-API, no login. Works across 8 Amazon marketplaces.

How many reviews you get per ASIN. Amazon publishes roughly 5 to 13 reviews on a public product page and no longer serves a public paginated review URL, so that is the ceiling. Setting max reviews to 100 does not get you 100, and it does not cost you anything either: billing is strictly per review delivered.

Yield still varies run to run, because Amazon does not always send a complete page. Three things can come back instead of the product page: a short bot-wall bounce, a page cut off above the reviews section, and, for some products, a reviews section with nothing in it. The actor tells those apart and reports them separately in the run summary. None of them is charged.

Amazon's official Selling Partner API only exposes reviews for products in your own approved catalog, and it requires seller registration, developer approval, and role authorization to get there. You cannot read a competitor's reviews through it at all. This scraper reads the same public product-review pages any shopper sees. Point it at any ASIN and get the reviews.

Reliability posture: blocked, empty, or failed runs are never charged. You only pay for a review record that was actually delivered.

✅ No login | ✅ No SP-API approval | ✅ Pay only per review returned | ✅ MCP-ready for AI agents

Features

Any public ASIN. Read reviews for your own products or any competitor listing. 8 marketplaces supported. US, UK, DE, FR, JP, IN, CA, AU. Sort and star filters. Order by recent or helpful, filter to a single star band. Verified-purchase flag. Every record carries a boolean so you can post-filter genuine buyers. Rich review payload. Full body text, images, helpful votes, variant purchased, and author profile. International reviews included and labelled. Amazon's "reviews from other countries" block is returned too, tagged with reviewSource and reviewCountry, and can be switched off.

How it works

For each ASIN, the actor fetches the public product page over plain HTTP and reads the reviews Amazon renders into it. No browser, no login.

Amazon does not always send the page. It has two ways of not sending it, and the actor treats them differently. A "Continue shopping" bot wall is retried on a fresh IP, up to five times. A page that arrives cut off above the reviews section is retried too, but only twice on the standard pool before the ASIN moves to a residential IP, because each of those cut pages is still half a megabyte or more to download. A product whose reviews section is present and empty is not retried at all: that is the product's real state, so the actor fetches it once, reports it, and moves on.

Sort order and the star filter are applied to the reviews on that page. Amazon serves one fixed set of top reviews per product, so these order and narrow what came back rather than requesting a different set.

You are billed per review returned, at $5.70 per 1,000 reviews. A blocked page, a missing ASIN, or a filter that matches nothing returns no rows and costs nothing.

🧾 Input configuration

{
  "asins": ["B08N5KWB9H", "B0BDHX8Z63"],
  "marketplace": "amazon.com",
  "maxReviews": 13,
  "sortBy": "recent",
  "filterByStar": "all",
  "includeInternational": true,
  "residentialFallback": true,
  "proxyConfig": { "useApifyProxy": true }
}

📤 Output format

{
  "reviewId": "R2STG6ZAMJNMK5",
  "asin": "B08N5KWB9H",
  "marketplace": "amazon.com",
  "rating": 4,
  "title": "Great product but took a week to ship",
  "body": "Love the build quality and the battery life exceeded expectations. Shipping was slow but the product itself is excellent. Would buy again.",
  "date": "2026-01-15",
  "dateLocalized": "Reviewed in the United States on January 15, 2026",
  "authorName": "J. Smith",
  "authorUrl": "https://www.amazon.com/gp/profile/amzn1.account.XXXX",
  "verifiedPurchase": true,
  "helpfulVotes": 12,
  "images": ["https://images-na.ssl-images-amazon.com/images/I/71xyz.jpg"],
  "variant": "Color: Black | Size: Large",
  "reviewCountry": "United States",
  "reviewSource": "local",
  "productUrl": "https://www.amazon.com/dp/B08N5KWB9H",
  "scraped_at": "2026-07-10T09:15:00.000Z"
}

Every review record contains these fields:

Field	Description
🆔 reviewId	Amazon review ID.
🏷️ asin	Product ASIN.
🌐 marketplace	Marketplace domain (e.g. amazon.com).
⭐ rating	Star rating from 1 to 5.
📝 title	Review headline.
📄 body	Full review text.
📅 date	Review date in ISO-8601 (best effort).
🗓️ dateLocalized	Date exactly as shown on the page.
👤 authorName	Reviewer display name.
🔗 authorUrl	Link to the reviewer profile.
✅ verifiedPurchase	True if the Verified Purchase badge is shown.
👍 helpfulVotes	Number of helpful votes.
🖼️ images	Review photo URLs (array).
🎨 variant	Variant purchased (e.g. Color: Black).
🌎 reviewCountry	Country the review was written in.
🧭 reviewSource	local for this marketplace's own reviews, international for Amazon's "reviews from other countries" block.
🌍 productUrl	Product page the review was read from.
🕒 scraped_at	ISO-8601 timestamp when scraped.

💼 Common use cases

Reputation monitoring Schedule runs to track review volume, average rating, and sentiment for your own catalog. Catch rating drops early before they hit BSR.

Sentiment analysis and NLP Feed the clean body text into classifiers, topic models, or LLMs to surface feature requests and pain points at scale.

Competitor research Filter to one- and two-star reviews with the star filter to read the highest-signal complaints on rival ASINs. Ideal for product design and copy angles.

Training datasets Pull star-labelled natural language for fine-tuning or evaluation without hand-scraping HTML. Every record ships with a clean rating column.

AI review digests Pair with Claude via MCP to summarize what customers love and hate about any product. Live pulls, no stale export snapshots.

🚀 Getting started

1. Open Amazon Reviews Scraper in Apify Console, or call it via API or MCP.
2. Under ASINs, paste one or more product IDs (e.g. B08N5KWB9H).
3. Pick your Amazon marketplace (default amazon.com).
4. Set max reviews per ASIN, a sort order (recent or helpful), and an optional star-rating filter. Expect 5 to 13 reviews per ASIN whatever cap you set.
5. Click Save & Start, then download the dataset as JSON, CSV, or Excel, or pull via the Apify API.

FAQ

Do I need an Amazon account or the SP-API? No. The scraper reads public product-review pages, the same data any shopper sees. No seller account, no SP-API approval, no login.

How many reviews can I scrape per ASIN? Roughly 5 to 13. That is what Amazon publishes on a public product page, and there is no longer a public paginated review URL behind it, so deeper history is not reachable without a logged-in session. The max reviews field caps the result but cannot raise that ceiling. Since billing is per review delivered, asking for more than Amazon has costs nothing extra.

Why did this actor used to promise more? It scraped /product-reviews/{ASIN} and paginated. Amazon now redirects that URL to its sign-in page, which is why the actor returned zero rows until this rebuild. The product page is the remaining public source.

Can I filter reviews by rating? Yes. The star-rating filter limits results to a single rating band (5-star through 1-star) or all. Every record also returns a verifiedPurchase flag so you can post-filter for verified reviews.

Do I need residential proxies? The residential fallback is on by default and you should leave it on. Amazon walls the shared datacenter pool hard. In a verification run on 20 August 2026, three ASINs came back as a page Amazon had cut short before the reviews section, and all three returned full pages when retried on residential. The actor starts on the cheaper pool and moves an ASIN to residential when it is blocked or when Amazon returns a page it cut short before the reviews section. If you would rather stay on the cheap pool and accept the lower yield, turn the fallback off.

How is it priced? Pay-per-result. You are charged per review returned, at $5.70 per 1,000 reviews. Rows that were never delivered are never charged, so blocked pages, missing ASINs, and empty filters cost nothing.

Use in Claude, ChatGPT & any MCP agent

https://mcp.apify.com/?tools=themineworks/amazon-reviews

Or call it from code with the Apify client:

import { ApifyClient } from 'apify-client';

const client = new ApifyClient({ token: 'YOUR_APIFY_TOKEN' });

const run = await client.actor('themineworks/amazon-reviews').call({
  asins: ['B08N5KWB9H'],
  marketplace: 'amazon.com',
  maxReviews: 13,
  sortBy: 'recent',
  filterByStar: 'all',
});

const { items } = await client.dataset(run.defaultDatasetId).listItems();
console.log(items);

🛠️ Complete your Amazon research pipeline

Pair with the rest of the suite for a full Amazon intelligence stack:

• Amazon Products Scraper: price, BSR, rating, and variants by ASIN or keyword.
• YouTube Channel Scraper: subscribers, video list, and channel stats, no API key.
• YouTube Transcript Scraper: timestamped captions as clean JSON for RAG and LLMs.

Found a bug or have a feature request? Open an issue on the actor's Apify Console page or reach out through the Apify profile.

Related guides

• Amazon Reviews Scraper: Collect Customer Ratings and Sentiment Without the SP-API

Last verified: 2026-08