Submit a Glassdoor company reviews link and get the reviews back as a table - the rating, the pros and the cons as separate columns, the reviewer's employment status, whether they recommend the company, and how they rated the CEO and the outlook.
That is an unusual thing for a product page to admit, and it is the most important sentence on this one.
Twelve runs of this service sit in front of us and every single one came back empty. What they do carry is a header, and that header is identical in all twelve: sixteen columns, listed in what you get back. So we can tell you the shape an export is meant to have, and we cannot show you a filled cell.
The distinction is worth being precise about. A header states what an export intends to contain. It is not evidence that a given column ever holds a value - a column can be declared and never fill. So nothing below is described as verified, and there are no fill rates on this page, because we have nothing to count.
The published reference does not close the gap either. It documents the request side thoroughly - what to submit, how many reviews per company, how to sort, how to cut off old ones - but it lists no response columns at all. That means nothing contradicts the sixteen names, and equally nothing independently confirms them.
Four settings, all of them documented, and two that decide how big a run gets.
Reviews per company defaults to 100, and the sort defaults to date. There is also a cut-off setting that takes an oldest timestamp, so a repeat run need not re-fetch the archive.
This is the header row of Livescraper reviews exports, identical across all twelve runs, in sheet order. It is not taken from the API reference - that document covers the request side and lists no result columns. Read it as the shape an export declares, not as columns we have watched fill.
Sixteen names, no values. The header is stable - identical across twelve separate runs - but every one of those runs came back without data rows, so we can tell you which columns arrive and nothing about what lands in them. Not the date format, not whether recommends is a word or a flag, not how often helpful is anything but zero. One run against the one-time 500 free rows answers all of that in a minute.
The column values are unknown, but the controls are not. These four are documented, and they are what shape the size and cost of a run.
The input is a direct Glassdoor company reviews URL - the reference gives https://www.glassdoor.com/Reviews/Amazon-Reviews-E6036.htm as its example. So the company has to be resolved to its Glassdoor page first; a bare company name is not the documented input. Up to 1,000 links go in a single run.
The limit is per query, not per run, which is the number that actually decides your bill: a hundred companies at the default is ten thousand rows, not a hundred. Set it deliberately before a wide run rather than discovering the multiplication afterwards.
Two options, and they answer different questions. Date order is what you want for monitoring - newest first, so a repeat run surfaces what changed. Relevance is Glassdoor's own idea of which reviews matter most, which is the better sample if you are reading a company once rather than tracking it.
The reference documents a cut-off that takes an oldest timestamp, so a scheduled pull can ask for nothing older than the last one instead of fetching the whole history and discarding most of it. On a company with years of reviews that is the difference between a monthly job that costs pennies and one that does not.
What a company's own staff reviews are good for, once they are in a spreadsheet.
The star average on the profile page is one number standing in for hundreds of opinions. Exporting the reviews puts the pros and the cons in separate columns, which is the form you need to count themes rather than skim them - and to see whether the complaints changed after you changed something.
Pull a competitor alongside yourself and the interesting part is the language, not the score. What their staff list under cons is a fair guide to where their hiring is vulnerable - and to what candidates will be comparing you against in an interview.
Acquirers, partners and senior candidates all end up reading the same reviews by hand. Having them as rows, sorted by date, makes it quick to see whether a pattern is historic or current - which is usually the question that actually matters.
No subscription, no minimum, no per-seat licence. Your first 500 rows are on us - after that, pay-as-you-go.
Every new account, one-time. No credit card required. All scrapers unlocked - and on this service in particular, the free rows are how you find out what the sixteen columns actually contain, which is something this page cannot tell you.
Roughly $2 per 1,000 reviews. Watch the per-company limit rather than the company count: at the default of 100 reviews each, twenty companies is two thousand rows. The estimator prices the run before it starts.
Volume pricing, SLAs, dedicated workers and bespoke onboarding for teams watching employer reputation across a set of companies on a fixed cadence. Tell us your numbers and we'll quote.
LIVESCRAPER10 at checkout.What a company says about itself and what its staff say rarely match. Reading both is the point.
The same site's jobs search as ranked rows, with the advertised salary and its source. What a company is offering, next to what its staff say about working there.
Review data we have measured properly - 842 real rows - if you want to see the shape a populated reviews export takes before committing to this one.
Short answer: yes, and one caution - these reviews are written by people about their own employment, which is more sensitive than most of the data on this site.
Glassdoor reviews are public. Anyone can read them without an account, and collecting publicly visible information for research is long-established practice. As long as the data is publicly available and the process doesn't disrupt the site, there are no federal laws prohibiting it.
The care needed here is greater than for a price list or a job advert. A review describes a person's experience of working somewhere, and the export declares a reviewer column and an employment_status column. We have not seen what either contains, which is exactly why we would treat both as personal data until a real export proves otherwise. In a small company, a role plus a location plus a date can identify someone even with no name attached.
So the sensible defaults: hold what you need and no more, drop identifying columns on import if your analysis is about the employer rather than the individuals, and do not attempt to work out who wrote a review. Glassdoor's terms also restrict automated access, so this remains a terms question. We touch nothing behind a login, run no third-party trackers on the data layer, and your exports auto-delete after 30 days.
The questions we hear most. Anything else? Talk to us - humans, not bots, write the answers.
query, company, review_title, rating, pros, cons, reviewer, employment_status, location, date, recommends, ceo_approval, outlook, helpful, review_url and position. That is the header of real Livescraper reviews exports, identical across twelve runs - not a list from the API reference, which specifies no result columns. The caveat is the important part: all twelve of those runs came back with the header and no data rows, so the names are what the export declares and not columns we have watched fill.https://www.glassdoor.com/Reviews/Amazon-Reviews-E6036.htm. A bare company name is not the documented input, so resolve the company to its Glassdoor page first. You can submit up to 1,000 links in a single run.reviewer or employment_status value from this service and we are not going to guess what they hold. Treat both as personal data until your own export shows you otherwise, and drop them on import if your analysis is about the employer.Submit one Glassdoor company link and see a populated export for yourself. Your first 500 rows are free - on this service that is not a marketing line, it is the only way to answer the question this page leaves open.
A company's Glassdoor profile compresses hundreds of individual accounts into a single star average and a rotating handful of featured reviews. For anyone whose work depends on what employees actually said - people teams reading their own reputation, talent teams sizing up a competitor, acquirers and senior candidates doing diligence - that summary is the wrong shape. The Glassdoor Reviews Scraper is built to return the underlying reviews as a table instead: submit a company's Glassdoor reviews link, choose how many reviews per company and whether to sort by date or relevance, and export as CSV, JSON or Excel.
This page has to be unusually careful about one thing, so it is said at the top rather than buried. We hold twelve runs of this service and every one of them came back empty. What those runs do carry is a header, identical across all twelve, naming sixteen columns - query, company, review_title, rating, pros, cons, reviewer, employment_status, location, date, recommends, ceo_approval, outlook, helpful, review_url and position. That header tells you the shape an export declares. It is not evidence that any particular column fills, and this page contains no fill rates, no sample values and no claims about formats, because there is nothing to count. Where other pages here quote measurements from hundreds or thousands of rows, this one names columns and stops.
The reference does not close that gap. It documents the request side thoroughly and lists no response columns at all, which cuts both ways: nothing contradicts the sixteen names, and nothing independently confirms them either. What the reference does settle is the set of controls, and those are worth knowing because they decide the size of a run. The input is a direct Glassdoor company reviews URL rather than a company name, up to a thousand links per run. The per-company review limit defaults to 100 - and being per company rather than per run, it is the number that actually multiplies your bill. Sorting is by date or by relevance, defaulting to date. And a cut-off parameter takes an oldest timestamp, which is what makes a scheduled re-run cheap on a company with years of history behind it.
One further note, on care rather than capability. These reviews are written by people about their own employment, which puts them at the more sensitive end of what this site collects. The export declares a reviewer column and an employment_status column, and since we have not seen either populated we would treat both as personal data by default. In a small organisation a role, a location and a date together can identify someone even with no name present. The sensible handling is the ordinary one: keep only what your analysis needs, drop identifying columns on import when the question is about the employer rather than the individuals, and make no attempt to work out who wrote what. Exports auto-delete after 30 days. Start free: your first 500 rows cost nothing, need no credit card, and will tell you more about what these sixteen columns contain than we honestly can.