Amazon holds the most detailed public record of what people buy online, what they pay, and what they say once the package arrives. Getting that information into a usable shape is the hard part. Anyone who has tried to collect it alone knows how fast page layouts change, how often CAPTCHAs appear, and how quickly an IP address gets throttled. That is why a growing number of analysts, sellers, and machine learning teams skip the collection step and buy or download a ready-made Amazon dataset instead.
This guide ranks the best Amazon datasets available today, from marketplaces that sell billions of records to free research corpora you can load into a notebook this afternoon. Record counts, prices, formats, and refresh options come from what each provider publishes on its own pages at the time of writing. If you are comparing the best Amazon datasets and databases for pricing intelligence, review analysis, seller research, or model training, the table below gives the short version, and the sections after it give the details.
Quick Comparison of the Best Amazon Datasets
What Is an Amazon Dataset?
- Product datasets hold listing details such as titles, brands, prices, images, and variations.
- Review datasets hold review text, star ratings, authors, dates, and helpful votes, tied back to an ASIN.
- Seller datasets describe marketplace sellers, with ratings, feedback counts, return policies, and fulfillment methods.
- Search and best seller datasets record what ranks for a keyword or sits at the top of a category.
- Deal and price history datasets track discounts and price changes over time.
Why Buy an Amazon Dataset Instead of Scraping Amazon?
Scraping gives you fresh pages, but fresh comes with a bill. Amazon layouts vary by category and region, defenses change without notice, and traffic that looks automated runs into 403 Forbidden errors before it ever reaches a product page. Teams that scrape at volume usually end up managing residential proxies for realistic traffic, sometimes ISP proxies when they need stable addresses that still read as home connections, and occasionally a set of mobile proxy providers for the stubborn cases. All of that is engineering time spent before the first usable row appears.
A dataset moves the burden to the seller. You get the same record structure on every delivery, a documented schema, and a price you can estimate up front. The trade-off is freshness. A snapshot is a point in time, so it suits historical analysis, benchmarking, and model training, while hourly repricing still calls for a live scraper API. Plenty of teams run both: datasets for the long view, and a scraper for the few hundred ASINs that need daily checks.
How We Ranked the Best Amazon Datasets
Six things decided the order. Coverage came first: how many records, which marketplaces, and which data types. Freshness followed, meaning how often a dataset refreshes and whether live data is available alongside it. Delivery mattered for anyone building a pipeline, so file formats and destinations such as S3, Snowflake, or SFTP counted. Price transparency carried real weight, because a published rate lets you budget before you talk to sales. Compliance statements and testability finished the list: does the provider explain how data is collected, and can you inspect a free sample first?
The ranking is an editorial judgment about fit for typical buyers, not a lab benchmark. A free research corpus can be the right answer even when it sits lower on the list, and an enterprise vendor in first place may be wrong for a student project.
The 8 Best Amazon Datasets & Databases
1. Oxylabs
Oxylabs sells ready-made e-commerce datasets with Amazon and Walmart as its main sources, and it is the provider to look at first if you want someone else to design the dataset around your question. Proxyway has named it the best overall e-commerce dataset provider in its own comparison, and the offer fits that framing: a standard product dataset plus fully custom builds.
The standard Amazon product data covers product name, brand, price, number of sellers, review count, rating, and availability, along with category details. Delivery is flexible. Files arrive as JSON, CSV, and other formats, either over SFTP or straight into cloud storage such as AWS S3 or Google Cloud Storage, and you pick a one-time, monthly, quarterly, or biannual schedule.
The custom route is where Oxylabs separates itself. You describe the data you need, the team designs a collection approach on its own infrastructure, you receive a sample to judge quality, and delivery then follows the schedule you agree on. A custom schema, a custom delivery frequency, and a dedicated Slack channel come with it. Oxylabs states that its datasets are compliant with GDPR and CCPA, and it also runs a dedicated Amazon scraper API for anyone who needs live pages rather than a snapshot.
Pricing is the catch. The company lists its datasets as starting from $1,000 and works by quote, so there is no per-record rate to drop into a spreadsheet. Expect a sales conversation before you see real numbers.
2. Bright Data
Bright Data runs the largest self-serve Amazon dataset catalog on this list. Its Amazon page currently reports more than 1.9 billion records spread across seven datasets: products, reviews, sellers info, a global products set, best seller products, product search results, and a combined Amazon and Walmart set. Narrower slices cover photos, fashion, electronics, books, beauty, and Prime listings.
Pricing is public. A one-time purchase starts at $250 for 100,000 records, which works out to as little as $0.0025 per record, and volume tiers run from 100K up to 20M records. Subscriptions cut the price by refresh rate: 25 percent off for biannual updates, 50 percent for quarterly, and up to 80 percent for monthly. Smart Data Updates let you pay only for new or changed records, and an AI filter turns a plain-English request into dataset filters so you buy only the slice you need.
Files arrive as JSON, NDJSON, CSV, XLSX, or Parquet with optional gzip, and delivery options include Amazon S3, Google Cloud Storage, Azure, Snowflake, Pub/Sub, SFTP, a webhook, or email. Every dataset offers a free sample before you commit. On compliance, the company cites GDPR, CCPA, ISO 27001, and SOC 2.
The weak spot comes with the size. More choices mean more filtering work, and the largest snapshots need real storage and pipeline capacity on your side. If you also need live data, its Amazon scraper API runs alongside the datasets.
3. Actowiz Solutions
Actowiz Solutions packages Amazon data into several focused datasets instead of one giant file. Its product, pricing, and review dataset page advertises coverage across more than 20 Amazon marketplaces, including the USA, UK, Germany, India, and Japan, with both real-time and historical data.
Field depth is the main reason to shortlist it. A product record can carry more than 50 fields, among them ASIN, title, brand, category, price, list price, deal price, coupon, Subscribe and Save price, Buy Box winner, seller count, rating, reviews, best seller rank, Prime status, stock, delivery estimate, seller name, and image URLs. For a pricing team, the Buy Box and deal fields can matter more than raw record count, because they show who actually wins the sale at a given moment.
The wider catalog adds seller data with fulfillment types such as FBA and FBM, a customer review set with sentiment classification, bestseller and trend data, and an inventory and stock set that tracks availability and regional delivery differences. Exports come as CSV, JSON, or Parquet, and updates can run daily, weekly, or monthly. Sample datasets are available for testing.
Actowiz does not publish prices, so plan for a demo call. It also does not list compliance certifications the way the larger vendors do, which matters if your legal team asks for them.
4. Axesso
Axesso Data Service is a European provider that started with an Amazon API and now sells Amazon data as downloadable datasets as well. Its dataset pages list a product dataset with more than 250 million entries and 25+ fields, a search results dataset with over 100 million results and 15+ fields, a review dataset with hundreds of millions of entries across 20+ marketplaces and a schema of 40+ fields, plus seller and deals datasets.
Refresh is its strength. Product, search, and review sets are updated weekly with several million new entries each time, and the deals dataset is updated several times a day across Amazon domains such as .com, .de, .fr, and .it. Files come as gzip CSV, JSON, or Excel, and data can also be shared through Google Analytics Hub, which gives subscribers direct SQL access to the underlying BigQuery table.
Free samples are available, and Axesso publishes an Amazon deals sample on Kaggle with more than 60,000 deals from amazon.com, so you can inspect its data style before you talk to anyone. Pricing follows record volume and dataset type rather than a flat fee.
Two limits are worth knowing. Compliance certifications are not listed on the pages we checked, and the company is narrower than the marketplace vendors, since Amazon is its core focus. For a team that lives in Amazon data and wants weekly freshness at a predictable per-record rate, that focus is a point in its favor.
5. Amazon Reviews 2023 (McAuley Lab on Hugging Face)
The strongest free option is Amazon Reviews 2023, built by the McAuley Lab and hosted on Hugging Face. The collection holds more than 570 million reviews and 48 million products across 33 categories, with interactions running through September 2023. Beyond review text and ratings, it includes helpful votes, product metadata such as descriptions, prices, and images, and user-to-product links that make recommendation research possible. It is reported to be 245 percent larger than the previous release, and it ships with standard train, validation, and test splits for benchmarking.
Treat it as a research resource. The data stops in 2023, which rules it out for current pricing work, and you should read the license terms before any commercial use. For model training, retrieval experiments, sentiment analysis, and recommender benchmarks, nothing free comes close in scale. The files are organized by category, and the benchmark folder alone is 366 GB, so download only the categories you need instead of pulling everything.
6. Apify
Apify runs a marketplace of ready-made scrapers called Actors, and its Amazon section works like a dataset factory. You pick an Amazon product, search, review, or seller Actor, feed it keywords, URLs, or ASINs, and the results land in an Apify dataset you can export as JSON, CSV, or Excel. Several Actors are priced by result, and the Store listings we checked ranged from roughly $3 to $7.50 per 1,000 results.
That model fits small and mid-sized jobs well: a few thousand competitor ASINs, a keyword research pull, or a weekly review export. Some product Actors cap a single run at 1,000 records, so larger collections mean chaining runs together. The bigger caution is maintenance. Most Amazon Actors are community-built; one search Actor we came across was flagged as under maintenance, and pricing can change without notice, so check the status and last update date on an Actor page before you build a pipeline around it.
You also get something datasets rarely give: control. A custom keyword list, a specific zip code for localized prices, or a single regional domain such as amazon.de are all inputs you set yourself. That flexibility is why Apify works best as a complement to a purchased dataset, filling the gaps a fixed snapshot leaves open.
7. Jungle Scout
Jungle Scout is closer to an Amazon market intelligence platform than a file download, but its data answers many of the same questions. It splits into two products. Catalyst serves smaller and growing sellers who want product-level research, with some features limited to the U.S. Cobalt targets established brands, agencies, and data teams that need a full market view, including first-party versus third-party sales analysis and pricing insights. Cobalt supports 19 Amazon marketplaces, and an API lets teams pull Jungle Scout data into their own reports and tools.
The best uses are sizing a category, estimating what competitors sell, and tracking brand share over time. It is the wrong pick if you want a bulk export of raw listings or reviews to feed a model. Keep in mind too that Jungle Scout's sales figures are estimates, so treat them as modeled numbers rather than Amazon's own.
Pricing splits the same way as the products. Catalyst runs on public plans with no contract beyond the current billing cycle, while Cobalt is sold through a demo and does not publish prices. Smaller sellers will likely find Catalyst enough, and brands with a large catalog will want to see Cobalt in action before deciding.
8. Datarade
Datarade does not own the data. It is a marketplace where providers list their Amazon data products, and buyers compare them side by side. The Amazon category currently lists 37 data providers, and there are 33 listings under Amazon reviews datasets alone. Names that show up include Webautomation, OpenWeb Ninja, Grepsr, PromptCloud, DataForSEO, TagX, and Bright Data itself.
The benefit is breadth. You can filter by data type, post a request, and let providers come to you, and many listings carry a 10 percent Datarade discount. The cost is inconsistency. Coverage, update frequency, compliance, and pricing all vary by seller, so a listing title tells you little until you request a sample, which Datarade itself recommends doing before any purchase.
Use it when your needs are unusual: an Amazon seller contact list, a regional extract, or a dataset built around a niche category that the large vendors do not sell off the shelf. For a standard product or review pull, going straight to a vendor with published pricing is usually faster.
How to Choose the Right Amazon Dataset
The right pick depends on the question you are asking, so start there.
For price monitoring, refresh rate matters more than size. Look at Actowiz for its Buy Box and deal fields, Axesso for a deals dataset updated several times a day, or a Bright Data subscription if monthly deliveries at a deep discount suit your budget.
For model training and retrieval projects, scale and a clean schema come first. Bright Data covers paid, current data in standard machine learning formats, while Amazon Reviews 2023 covers free historical text and metadata.
For review analysis, compare the review schemas directly. Axesso lists a review dataset with 40+ fields, Bright Data sells a dedicated reviews set, and the McAuley Lab corpus is the standard free benchmark.
For seller research, check which seller fields you actually receive. Bright Data and Axesso both sell seller datasets, and Jungle Scout covers the market side with brand and category views.
On a tight budget, start with the free Amazon Reviews 2023 corpus or a small Apify run, then upgrade when the project proves its value. Sellers who also compete on other marketplaces can apply the same logic elsewhere, and a look at Etsy proxies shows how the collection side differs on a different platform.
Before any purchase, run the same short checklist. Request a sample and check column names, null rates, and duplicate ASINs. Confirm which Amazon domains are included, since a dataset heavy on amazon.com may thin out for amazon.in or amazon.co.jp. Ask how often the data refreshes and whether you can receive only new or changed records. Read the licensing terms for commercial use, especially on free research datasets.
Finally, ask how the provider describes its collection method. Bright Data says its data comes from publicly available sources, and Oxylabs states GDPR and CCPA compliance, but your legal team may still want those statements in writing. None of this is legal advice, and any project that touches personal data deserves a professional review.
Amazon Dataset FAQs
Where can I find free Amazon datasets?
Amazon Reviews 2023 on Hugging Face is the main free source, and Axesso's Kaggle deals sample works for testing a schema. Both are snapshots, so confirm the collection date before using one for current pricing.
How much does an Amazon dataset cost?
Prices run from free to enterprise level. Bright Data starts at $250 for 100,000 records, Oxylabs lists datasets from $1,000, and Apify Actors charge roughly $3 to $7.50 per 1,000 results. Subscriptions lower the per-record cost, and Bright Data discounts its monthly refresh plan by up to 80 percent compared with a one-time purchase.
How often should an Amazon dataset refresh?
Match the schedule to how fast your question changes. Price and deal tracking benefits from daily or weekly updates, while market sizing and model training usually work fine on monthly or quarterly snapshots. Vendors such as Bright Data, Oxylabs, and Axesso offer schedules that range from a one-time purchase to weekly refreshes, and some go further.
Is it legal to use an Amazon dataset?
Datasets built from publicly available pages are widely sold, and several vendors state that they comply with GDPR and CCPA. Legality still depends on your jurisdiction, how the data was collected, what it contains, and how you use it. Ask the provider for compliance documentation, and speak with a lawyer before commercial use or any work involving personal data.
Should I buy a dataset or use a scraper API?
Buy a dataset when you need history, scale, or a clean file for analysis or training. Use a scraper API when you need live prices or stock for a defined list of ASINs. Many teams combine the two, with a dataset for context and an API for the items that change by the hour.
A practical first step is to download free samples from two vendors that fit your use case, load both into the same notebook, and compare how many ASINs overlap and which fields are filled. That short test tells you more than any feature list.
.png)
No comments