Paste Shopee product URLs, one per line, and the review section comes back as a table - who wrote it, the rating, the comment, the date, how many people found it helpful, which variation they actually bought, the product, and any images attached. Nine columns, one row per review.
A list of product URLs and two controls. Everything else is the export.
…-i.<shopid>.<itemid> form. Or upload a CSV, XLSX, TXT or Parquet file.The limit is per query, not per run: five product URLs with a limit of two hundred is two hundred reviews from each, not two hundred altogether.
The sorting control decides what comes back: everything, only reviews that carry a written comment, or only the ones with photos or video attached. Because you are billed per row, narrowing to With comment on a product with thousands of bare star ratings is the difference between a useful file and an expensive one.
A marketplace listing is rarely one product. variation records which option the reviewer actually bought, which is what lets a complaint be attributed to a size, a colour or a bundle rather than to the listing as a whole. That is usually where the real signal is hiding.
The column set does not change between runs, so the same product URLs a month apart give two files that line up column for column. That is what makes a diff possible at all - and query is echoed on every row, so a file covering several products still splits cleanly.
These names are confirmed twice: they are the service's own declared list, and they are the header row of every run export we hold. What each column is FOR is below. What each column will CONTAIN is not - the note explains why we will not guess.
The names are solid. Everything past the names is your first run's job, and that is a deliberate limit rather than an omission. We hold five run exports for this service and every single one came back empty - each JSON file is literally an empty list, and each spreadsheet contains exactly one row: the header. That header is byte-identical across all five and matches the service's own declared column list, which is why the names above can be trusted completely. Nothing else can. So this page prints no example values, no formats, no units, no types and no fill rates - not for rating, not for date, not for images. We could copy plausible-looking answers off Shopee's website or off another reviews page here, and they would be invention. The reason the runs came back empty is the one the service names itself: Shopee geo-blocks datacenter traffic and needs a regional residential proxy. Run the free tier against one product and read the header row and the first few values before you build anything that parses them.
All of them start from a product link you can copy out of the app.
An average tells you something is wrong; it never tells you what. Filter to With comment, read the low ratings against variation, and the pattern usually resolves into one option - a size that runs small, a colour that arrives wrong, a bundle missing a part.
A competitor's review section is the most candid description of their product you will ever get, and it is public. Point the scraper at their listing and you get the same nine columns you get for your own, which means the two files compare directly.
The With media filter narrows to reviews carrying images. For a defect investigation or a supplier conversation, a handful of photographed complaints is worth more than a thousand bare star ratings - and it keeps the row count, and the bill, small.
No subscription, no minimum, no per-seat licence. Your first 500 rows are on us - after that it is pay-as-you-go.
For every new account, one time. No credit card. All scrapers unlocked. Given that no run we hold has returned data, this is the part that matters most here: 500 rows is enough to settle for yourself what this export actually contains before you plan around it.
Roughly $2 per 1,000 reviews, the same flat rate as every other scraper on the platform. The limit and the sorting filter are the two dials that keep a busy listing in proportion - narrowing to reviews with comments is both cheaper and, usually, more useful.
Volume pricing, SLAs, dedicated workers and tailored onboarding for teams tracking large catalogues across marketplaces. Tell us your numbers and we will quote.
LIVESCRAPER10 at checkout.Reviews live wherever the product is sold. These return them as rows too.
The same job on the other marketplace most of these products are also listed on. Useful when you want to know whether a complaint is about the item or about one seller's handling of it.
Reviews of the seller rather than the item. A marketplace listing and the company behind it collect two different kinds of complaint, and both are worth reading.
If the shopping happens in an app, the app's own reviews are a separate and often blunter channel. Same shape of output: one row per review.
Short answer: yes, when you only collect what is publicly visible on a product page - and that is all this service reads.
A product's review section is public. Anyone can open the listing without an account and read it, and collecting publicly visible information for product and market research is long-settled practice. As long as the data is publicly available and the process does not disrupt the service, there are no federal laws against it.
Reviews are written by people, though, and author is a person's display name. Publicly visible is not the same as free of obligation: if you store it, you are handling personal data, and the GDPR and similar regimes apply to you regardless of where you got it - as do the data-protection laws of the market the listing sits in, which for Shopee usually means somewhere in Southeast Asia. Reading complaints to fix a product is an easy case; building a durable file on individual reviewers is not, and this service is not intended for it.
Shopee's terms restrict automated access, so this remains a question of terms. We touch nothing behind a login, read only what an ordinary visitor sees, run no third-party trackers on the data layer, and your exports self-delete after 30 days.
The questions we hear most. Something else? Talk to us - humans write the answers, not bots.
Paste one Shopee product URL, pick a filter, and see what comes back. Your first 500 rows are free - and with this one, that first run is genuinely the point.
A marketplace review section is the most candid description of a product anyone will write, and it is almost unusable in the browser: you can scroll it, and that is all. You cannot sort it your own way, filter it to the complaints, or line one listing up against another. This service turns it into a table. Paste Shopee product URLs, one per line, and the reviews come back as rows - the reviewer, the rating, the written comment, the date, how many people found it helpful, which variation was bought, the product, and any images attached. Nine columns, one row per review.
There are two controls and both earn their place. The limit is per query rather than per run, so five product URLs with a limit of two hundred returns two hundred reviews from each rather than two hundred in total - set it to 0 to take everything a listing holds. The sorting control decides what counts as a review worth returning: All, With comment, or With media. On a listing carrying thousands of bare star ratings, narrowing to the ones with actual text is both cheaper - you are billed per row - and closer to what you were looking for anyway.
The column worth planning around is variation. A marketplace listing is rarely a single product: it is a size, a colour, a bundle and a two-pack sharing one page and one average rating. Recording which option the reviewer actually bought is what turns a vague one-star into an attributable defect, and it is why a complaint that looks like it is about the product often turns out to be about one variant of it.
One limit is stated plainly on this page because it changes what you should expect from it. We hold five run exports for this service and every one came back empty - each JSON file is an empty list and each spreadsheet is a header row with nothing under it. The reason is the one the service gives itself: Shopee geo-blocks datacenter IPs and needs a regional residential proxy. What survives from those runs is the column names, and they survive well: the header row is identical across all five workbooks and matches the service's own declared list. So the names above are dependable, and everything past them - formats, units, what a rating looks like, what a date looks like, whether images arrive as one value or several - is deliberately absent rather than guessed. Run the free tier against a single product and read the first few rows yourself; your first 500 cost nothing and need no credit card. See pricing for current rates.