Amazon reviews dataset in 2026
Prices compared, products as rows.
Bright Data sells its Amazon Reviews dataset, 43.1M records, at up to $0.0025 a record, from $250 for 100K records. Apify's Amazon Reviews Scraper starts at $3.00 per 1,000 reviews. McAuley Lab's 571.54M reviews end in September 2023.
SpecStream is the live web data engine for teams and agents. Describe the products you track, and Deep Search returns each one as a row with its rating, rating count and published reviews, at 5¢, 15¢ or $1, rerun as often as hourly.
New accounts get $10 of credit, no card.
Amazon review data sources side by side
| SpecStream | Bright Data Amazon Reviews dataset | Apify Amazon Reviews Scraper | McAuley Lab Amazon Reviews'23 | |
|---|---|---|---|---|
| What you ask for | The products you track in plain words, with the conditions every row must meet: category, store, price, star rating, rating count | A pick from its Amazon datasets, narrowed by filters or by one prompt in plain English | Amazon product URLs, with an optional cap on the number of reviews | Nothing to ask: one review file and one metadata file per category, 33 categories |
| What a row carries | One product per row: the facts you name, such as brand, price, star rating, rating count and the reviews its page publishes, joined into one record. Source URL and read time on every value | One review per record, 27 data fields: product name, rating, review text, headline, author name, ASIN, posting date, helpful count, review images and more | One review per result, with rating score, review description, reactions and images. Only ratings that carry a text review, at most 100 per star rating, so 500 per product at best | One review per line, with rating, text and helpfulness votes, plus item metadata such as description, price and images |
| How old the data is | Read from the live web during your run, with the read time on every value | A snapshot of pre-collected records, refreshed monthly | Read from Amazon when you run it | Reviews from May 1996 to September 2023, collected in 2023 |
| The unit you pay for | A delivered row: 5¢, 15¢ or $1 by how much the row needs. A product with up to 15 facts and its reviews is a 15¢ row. Only rows that meet every condition are billed | A record, at up to $0.0025, which is $2.50 per 1,000 reviews | A review, from $3.00 per 1,000 | A direct download link for each category file |
| Smallest start | $10 of credit on sign-up, no card. A top-up starts at $10 | $250, one time, for the 100K record tier | A free trial on the Apify Store, then pay per review | One category file, such as All_Beauty with 701.5K ratings |
| A custom dataset | Every request is a custom dataset: the products, stores and facts you write, priced at the same row rates and quoted before it runs | One prompt sets the filters: "Describe exactly what you need, and let AI apply the perfect filters in seconds." A subset with only the data points you need costs less | The products you list by URL, one run at a time | The categories it collected, with 0-core and 5-core subsets |
| Other Amazon data | Products, sellers and best sellers are requests too: name the category, the store and the facts, and get one row per match | Separate datasets: Amazon products 758.5M records, best seller products 171.7M, products search 20M, sellers info 3.1M | Separate actors, such as an Amazon product scraper for price, Best Sellers Rank and review count | Item metadata for 48.19M items in the same release |
| Scheduled delivery | A saved request reruns hourly to yearly, inside the spending limits you set | Updates daily, weekly, monthly or on a custom basis. A biannual, quarterly or monthly plan saves 25%, 50% or 80% | Runs you schedule through the Apify API | One release, with no updates after September 2023 |
| Only what changed | Each row comes back labeled changed, unchanged or no longer matching. A changed or new row costs half its rate, an unchanged row 1¢ | Smart data updates send only "New Records" or "Updated Records", and you pay for those records | Each run reads the reviews again, and a Max Reviews cap shortens the run | A new release replaces the last one: the 2018 release held 233.10M reviews |
| Where the data goes | Google Sheets, Excel, Airtable, Notion, monday.com, HubSpot, Salesforce, Pipedrive, Attio, Slack, Teams, Zapier, Make, n8n and signed webhooks | JSON, NDJSON, JSON Lines, CSV or Parquet files, sent to Snowflake, Amazon S3, Google Cloud, Azure or SFTP | Downloads as HTML table, JSON, CSV, Excel or XML | A review file and a metadata file for each category |
| MCP and API | MCP server, REST API, Python and TypeScript SDKs, a CLI and A2A. An agent can buy one request with no account, by card or in USDC | API delivery of the dataset into your own systems | The Apify API, to run actors and read their datasets | The Hugging Face datasets library |
Why Amazon review data buyers move to SpecStream
The products you track, with their reviews
A review dataset sells one record per review, so the question you start from, which products in my category rate how and why, is a join you write yourself. SpecStream starts from that question. Write the products you track, such as "air fryers sold on Amazon US with at least 1,000 ratings", and each row is one product with the facts you named and the reviews its page publishes, joined into one record. Bright Data's prompt sets filters on the fields its dataset already holds. On SpecStream a condition can be any trait a page states, and each row shows the page that states it.
Start with $10, not a $250 order
The smallest Bright Data order is $250 for 100K records. On SpecStream you get $10 of credit on sign-up, no card, and pay for the rows that meet your request: 5¢ for a row with up to 5 facts, 15¢ for up to 15 facts and one set of related records, such as a product's reviews, and $1 for a deep row with no limit. At 15¢, $10 covers 66 products with their reviews. The most a run can cost is quoted before it starts, and rows you already own are never charged again.
Schedules that label every change
Bright Data updates the Amazon Reviews dataset daily, weekly, monthly or on a custom basis, and its smart data updates send only new or updated records at the same price per record. A saved SpecStream request reruns hourly to yearly, and each row comes back labeled changed, unchanged or no longer matching, so a product whose rating dropped comes back labeled changed. A changed or new row costs half its rate (3¢, 8¢ or 50¢), and an unchanged row 1¢. The rows land in Slack, email, Sheets, HubSpot, Salesforce or a signed webhook, or in your agent through the MCP server.
Today's ratings, not 2023's
McAuley Lab's Amazon Reviews'23 holds 571.54M reviews from May 1996 to September 2023, and the release before it, in 2018, held 233.10M. A release is a fixed copy: a product launched since then is not in it, and a rating that moved since then shows its old value. SpecStream reads each product's page during your run and writes the read time beside every value, so the rating in your sheet is the rating a shopper sees today.
Other Amazon reviews dataset options
Outscraper Amazon Reviews Scraper
A reviews scraper billed by monthly usage: the first 500 reviews are free, then $2 per 1,000 up to 5,000 reviews and $1 per 1,000 after. SpecStream returns the product with its rating, rating count and reviews as one row, from $10 of credit.
Axesso Amazon Reviews Scraper on Apify
An Apify actor at $0.90 per 1,000 reviews that returns ratings, titles, review text, reactions, dates and images. You give it the products. SpecStream finds the products that meet your conditions, then reads their reviews into the same row.
Bright Data Amazon Reviews Scraper API
Bright Data's API for live review records, with 5K free records a month and no card required, billed per 1,000 records after that. SpecStream bills per delivered row and quotes the most a run can cost before it starts.
Datarade
A data marketplace that lists 18 Amazon reviews dataset products from different providers, each priced by its seller. SpecStream gives one price for every request: 5¢, 15¢ or $1 a row, the same on every plan.
How SpecStream scores
The same numbers as the benchmarks page, from the official tasks and scorers.
| SpecStream |  Parallel |  Exa |  Tavily |  Perplexity | |
|---|---|---|---|---|---|
| Web researchWideSearch, GPT-5.6 Luna | 60.9 | 45.5 | 53.0 | 47.9 | 47.0 |
| Factual accuracySimpleQA Verified, GPT-5.6 Luna | 96 | 94 | 91 | 94 | 94 |
| Factual accuracySimpleQA Verified, GPT-5.6 Sol | 99 | 97 | 91 | 92 | 95 |
How to switch from an Amazon reviews dataset
Write the products you track in plain words, in place of the dataset filters you would set. Sign up and you get $10 of credit, no card. At 15¢ a row, that is 66 products with their reviews.
Or send the request from code. This call spends nothing: the 402 reply shows how SpecStream read the request, the columns and the most it can cost.
curl -i https://api.specstream.ai/v1/agent/lists \
-H "Content-Type: application/json" \
-d '{"request": "Air fryers sold on Amazon US with at least 1,000 ratings, with the brand, the price, the star rating, the rating count and the reviews shown on each product page", "count": 25}'Approve that maximum by card through Stripe's Link agent wallet, or in USDC, and you pay for the rows found. On an account, your agent uses the MCP server at https://api.specstream.ai/mcp and your team uses the workspace, from the same balance. To keep the rows up to date, save the request and choose how often it reruns.
Comparing all of Bright Data's products? See Bright Data alternatives.
Frequently asked questions
Something we didn't answer? Talk to a human
How much does the Bright Data Amazon reviews dataset cost?
What is the best Amazon reviews dataset alternative?
How many reviews can I scrape per Amazon product?
What does the McAuley Lab Amazon reviews dataset hold?
Can I get only new or updated Amazon reviews?
Does Bright Data build custom Amazon datasets?
Does SpecStream have an MCP server?
https://api.specstream.ai/mcp over Streamable HTTP, with OAuth or an API key. Connecting an agent is free on every plan, and it spends from the same balance as your team, inside the limits you set on its key.Sources
SpecStream wrote this page. Each rival figure comes from that company's own page.
- SpecStream pricingspecstream.ai/pricing, read Nov 12, 2026
- SpecStream for agentsspecstream.ai/agents.md, read Nov 12, 2026
- Bright Data Amazon Reviews datasetbrightdata.com/products/datasets/amazon/reviews, read Nov 12, 2026
- Bright Data datasetsbrightdata.com/products/datasets, read Nov 12, 2026
- Bright Data Scraper Studio pricingbrightdata.com/pricing/datasets/custom, read Nov 12, 2026
- Bright Data Amazon Reviews Scraper APIbrightdata.com/products/web-scraper/amazon/reviews, read Nov 12, 2026
- McAuley Lab Amazon Reviews'23huggingface.co/datasets/McAuley-Lab/Amazon-Reviews-2023, read Nov 12, 2026
- Apify Amazon Reviews Scraperapify.com/junglee/amazon-reviews-scraper, read Nov 12, 2026
- Axesso Amazon Reviews Scraperapify.com/axesso_data/amazon-reviews-scraper, read Nov 12, 2026
- Outscraper Amazon Reviews Scraperoutscraper.com/amazon-reviews-scraper/, read Nov 12, 2026
- Datarade Amazon reviews datasetsdatarade.ai/search/products/amazon-reviews-dataset, read Nov 12, 2026