What is Amazon Review Scraper?
Amazon Review Scraper reads the customer reviews of one Amazon product, identified by its ASIN or link, and returns each review as a record. Every record also carries the product's average rating and its counts of ratings and written reviews, so you know how much of the picture your sample covers.
You choose which reviews to read: the most helpful or the most recent, all stars or only the critical ones, every variant or one color and size.
You ask
Pull the most helpful Amazon reviews for the AirPods Pro 2 (B0D1XD1ZV3) with rating, title, helpful votes and verified purchase.
It runs
From your AI assistant over MCP or from your code over REST
You get
- Review
- Buyer
- Photos and video
- Product
- Request
- Check a feature before you stock itRead what owners say about battery life, fit or noise before a product goes into your catalog or a buying guide.
- Mine a rival's complaintsPull the critical reviews of a competing product and list the defects buyers report most.
- Fix your own listingFind the doubts buyers raise about size, setup or compatibility and answer them in your bullets and images.
- Quote real ownersUse verified-purchase reviews, with dates and helpful votes, in roundups and video scripts.
- Let your agent read for youAsk Claude or ChatGPT to read 100 reviews and sort praise from complaints in the same chat.
What data does the Amazon review scraper extract?
Each review comes back as one record with the product context attached. These are the fields from our test run.
Review
- Star rating
- Review title
- Full review text
- Date and country of the review
- Review ID
- Helpful votes
Buyer
- Reviewer name and profile link
- Verified purchase badge
- Vine program flag
Photos and video
- Links to photos in the review
- Links to videos in the review
Product
- ASIN and product title
- Variant bought, such as style, color or size
- ASIN of that variant
- Average rating
- Total ratings and written reviews
Request
- Marketplace
- Sort order
- Page number
- Filters applied
Reviewer names and profile links are personal data, so this page masks them; keep them out of your dataset unless you need them. For price, bullets and images of the product itself, use Amazon Product by ASIN.
Sample output: AirPods Pro 2 reviews
The table at the top is a real run from 16 November 2026: the first page of amazon.com reviews for Apple AirPods Pro 2 (ASIN B0D1XD1ZV3) in Amazon's most-helpful order. It returned 10 reviews. The product has a 4.6 average from 47,174 ratings, 8,330 of them with written reviews. Below is the one review in the run under five stars.
One result in full
One review in full
- Product
- Apple AirPods Pro 2 (ASIN B0D1XD1ZV3)
- Rating
- 4.0 out of 5 stars
- Title
- Big Improvement over 1st Gen, not so sure they are hearing aids
- Text
- I was fairly satisfied with my first generation AirPods Pro, having bought them several years ago mostly for use when working out. When Apple announced the addition of hearing aids features, I thought it was a good…
- Reviewer
- J•••• G. (profile link returned)
- Date and country
- Reviewed in the United States on November 20, 2024
- Verified purchase
- Yes
- Vine review
- No
- Variant
- Style: Without AppleCare+
- Helpful votes
- 69
- Photos and videos
- 0 and 0
- Product rating
- 4.6 out of 5 from 47,174 ratings, 8,330 with text
What this sample tells you
9of 10
9 of 10 reviews give five stars and one gives four, above the product's 4.6 average. The most-helpful order favors long, happy reviews, so filter by stars when you study complaints.
All 10 are verified purchases and none come from the Vine program. Helpful votes range from 0 to 881, and one review carries 5 photos and a video.
5of 10
These reviews are long: 149 to 917 words, 504 at the median. 5 of 10 discuss the hearing aid feature and 6 mention battery life, from “the only thing I’m dissatisfied with” to “fantastic”.
Dates run from August 2024 to October 2026, so one page mixes early owners with recent ones. Sort by most recent to see how opinions change over time.
How to scrape Amazon reviews with your AI assistant
Connect AI/ML API
Use Try it now to add the AI/ML API MCP server to your client. It takes a minute and one sign-in.
Point to the product
Paste the Amazon link or the ASIN. If you only know the name, your assistant can find it first with Amazon Product Search.
Choose which reviews
Most helpful or most recent, a star filter such as critical only, a keyword, and the marketplace, for example amazon.co.uk.
Turn reviews into answers
Ask for a table, a count of mentions per feature, quotes for and against, or a CSV file.
Prompts to copy
Use the AI/ML API MCP Amazon Review Scraper tool. Pull the 10 most helpful amazon.com reviews for ASIN B0D1XD1ZV3 and count how many mention battery life, fit or noise cancellation.
Use the AI/ML API MCP Amazon Review Scraper tool. Get the critical reviews, 1 to 3 stars, of the Stanley Quencher 40 oz tumbler and group the complaints by theme.
Use the AI/ML API MCP Amazon Review Scraper tool. Compare verified-purchase reviews of the Kindle Paperwhite on amazon.com and amazon.co.uk: average stars and the top three complaints in each.
Use the AI/ML API MCP Amazon Review Scraper tool. Find the 20 most recent amazon.com reviews of the Ninja Creami that mention “broken” or “stopped working” and list them by date.
Review filters and settings
Your assistant picks these from your request. In code you pass them yourself.
| Setting | What it controls | Tip |
|---|---|---|
| Product | ASIN or product link, one or several | The ASIN is the 10-character code after /dp/ in the link |
| Marketplace | amazon.com, amazon.co.uk, amazon.de, amazon.co.jp and 18 more | Reviews differ by country, so ask the market you sell in |
| Sort | Most helpful or most recent | Recent shows how a product holds up after changes |
| Star rating | All, one to five stars, positive or critical | Critical gives you the complaints in one pass |
| Keywords | Only reviews that contain your words | Use the buyer's words: “leak”, “battery”, “returned” |
| Verified only | Keep only verified-purchase reviews | Useful when you quote owners in an article |
| Variants | Reviews of every color and size, or only the variant you name | Turn it off when one color has a known defect |
| Pages and dates | Pages per product, 10 reviews per page, and an optional date range | A date range lets you compare before and after a product change |
Amazon shows at most 10 pages, about 100 reviews, for one combination of sort order and star filter. To reach more, request each star level separately: that returns up to 100 reviews per star and up to 500 per product. Not every filter is available on every request, so if one is ignored, ask your assistant to filter the returned reviews instead.
Ways sellers and researchers use Amazon reviews
Research before sourcing
Read the reviews of the best sellers in a category and note the defects you could fix in your own version.
RatingTextHelpful votes
Listing copy that answers doubts
Turn recurring buyer questions about size, setup or compatibility into bullets and images on your product page.
TextVariantRating
Competitor teardown
Pull a rival's critical reviews and rank the complaints by how often they appear.
RatingTextDate
Quality tracking after a change
Compare reviews before and after a new batch, supplier or firmware update with a date range.
DateRatingText
Buying guides and video scripts
Quote verified owners with dates in roundups, and note how many helpful votes each quote earned.
VerifiedHelpful votesDate
Datasets for models
Collect review text with its star rating as the label for sentiment or feature-level classification.
RatingTitleText
Amazon reviews scraper API for your code
In your code
REST · your AI/ML API key, JSON records back
The same tool runs over REST with your AI/ML API key. Send ASINs, a marketplace and your filters, and get back JSON with one object per review, each carrying the product's title, average rating and totals.
Schedule it for the products you sell or watch, store the review IDs and keep only the new ones on the next run. Use MCP for questions asked in a chat and REST for jobs that run on their own, both on one balance. See the Toolbox REST quickstart.
Other ways to get Amazon reviews
Amazon's own APIs are built for selling and affiliate links rather than for reading review text, so most teams copy reviews by hand or scrape them.
| Read and copy on Amazon | Amazon's affiliate or seller APIs | Your own scraper | ||
|---|---|---|---|---|
| Setup | Connect once with your AI/ML API account | None | Approved account, your own code | Proxies, sign-in handling and parsers to maintain |
| Inside an AI chat | Yes. Over MCP | No. No | Only if you build it | Only if you build it |
| Best for | Repeat review research from chat or code | A few reviews of one product | Prices, listings and orders, not review text | Teams with engineers to keep it running |
Pricing
Pay as you goBilled per run from your AI/ML API balance
- No subscription
- No seat fee
- One balance for every tool
Runs are billed from your AI/ML API balance, with no subscription or seat fee. Our test run, one page of 10 reviews, cost just over one cent. The cost grows with the number of reviews you collect, so check the current rate in your dashboard before you pull thousands.
FAQ
Is it legal to scrape Amazon reviews?
Reviews are public on product pages, and reading them for product research is common. Amazon's conditions of use apply to the site, and privacy laws such as GDPR apply to reviewer names. Keep personal data out of your dataset unless you need it, and ask a lawyer about your specific case.
How many reviews can I get per product?
Amazon shows up to 10 pages of 10 reviews for one sort order and star filter. Request each star level separately to reach up to 500 reviews per product.
Can I get only the negative reviews?
Yes. Choose critical or a single star level, and add keywords such as “broken” to narrow them down.
Does it work outside amazon.com?
Yes, on 22 Amazon marketplaces, including amazon.co.uk, amazon.de, amazon.fr, amazon.co.jp, amazon.in and amazon.com.br.
Does it detect fake reviews?
No. It returns the verified-purchase badge, the Vine flag, helpful votes and dates, which you or your assistant can use as signals. It does not score authenticity.
Can I get the product price and description too?
Each review carries the product title and rating totals. For price, bullets and images, run Amazon Product by ASIN on the same ASIN.
How is this different from Amazon's own APIs?
Amazon's affiliate and seller APIs serve product, listing and order data to approved accounts, not the text of customer reviews. This tool reads the reviews shown on the product page and works from AI assistants over MCP.
Can I use it from Claude, ChatGPT or Cursor?
Yes. Add the AI/ML API MCP server and paste a product link into the chat. Other MCP clients work the same way.
Can I export Amazon reviews to CSV?
Ask your assistant for a CSV, or write the REST response to a file or a spreadsheet in your code.
How fresh are the reviews?
Each run reads Amazon at that moment. In our most-helpful test the newest review was from 27 October 2026, three weeks before the run; sort by most recent to get this week's reviews.
Read your first Amazon reviews
Paste a product link and ask what buyers say about it.