Livescraper
Sign inGet Started
G2 Reviews Scraper

Pros and cons,
as separate columns.

Give it a G2 product URL - or just the slug - and its reviews come back as a table. G2 asks reviewers what they liked and what they disliked as two different questions, and this export keeps them that way: pros and cons are their own columns, next to the rating, the full review text, the reviewer's company size and the date.

Start free See pricing
one-time 500 free rows$0.002 per row afterone row per reviewCSV · JSON · Excel
How it works

A slug is enough
to start.

One required field and two controls, and you do not need the full URL for either.

  1. STEP 1Sign in to the platform.
  2. STEP 2Open the G2 Reviews Scraper.
  3. STEP 3Paste a G2 reviews URL or just the product slug, one per line - or upload a CSV, XLSX, TXT or Parquet file.
  4. STEP 4Set a limit per query, or leave it empty to fetch all. The field's minimum is 1, and it does not accept 0.
  5. STEP 5Pick a sort: the default G2 Sort, or Most recent, Most Helpful, Highest rated or Lowest rated.
  6. STEP 6Click Start scraping and download as CSV, JSON or Excel.

Both input forms are confirmed rather than assumed: we hold runs for the same product submitted once as a bare slug and once as a full URL, and both returned it.

Why teams use it

The structure is the point.
G2 already split the review.

Complaints you can read on their own

Because G2 asks for likes and dislikes separately, cons arrives as a column you can read end to end without wading through the praise around it. On most review sites you would have to infer the criticism from a paragraph; here it is already isolated, and it is the single most useful thing in the export.

Segment by who is complaining

company_size is a closed set - in our sample it was always one of Small-Business (50 or fewer emp.), Mid-Market (51-1000 emp.) or Enterprise (> 1000 emp.). A complaint from a fifteen-person team and a complaint from an enterprise buyer are different products' problems, and this is the column that separates them.

Sorted the way you need, not the way G2 defaults

The sort has five settings, and the default is G2's own ordering. Switching to Lowest rated is what gets you the reviews the page will not lead with - and it matters here more than usual, because the default ordering skews high enough that it shapes what you conclude.

Data dictionary

Sixteen columns,
three of which never filled.

The names are the header row of all twenty-seven exports we hold, identical across every one. The behaviour below was measured on fifty-nine reviews spanning four products - that sample is named again wherever it matters.

query
What you submitted for this row - a reviews URL or a bare product slug - echoed back on every row it produced.
product_name
The product the review belongs to. In our sample this arrived as G2's page heading rather than a bare name, in the form <Product> Reviews & Product Details, so strip the suffix if you want just the product.
product_url
The product's reviews page on G2.
review_title
The review's headline. Short - five to thirty-four characters across our fifty-nine.
rating
The review's rating, as a string rather than a number. Every value we saw was '5.0' or '4.5' - but that is what the default sort surfaced, so treat it as a fact about our sample rather than about G2's scale.
review_text
The full review body. Substantial on this site: six hundred to nearly five thousand characters in our sample.
pros
What the reviewer said they liked, as its own column - G2 asks this as a separate question.
cons
What the reviewer said they disliked, likewise its own column. This is the one to read first.
reviewer_name
The reviewer, as the review is credited.
reviewer_title
Presumably the reviewer's job title.
company_size
The reviewer's company size band. A closed vocabulary: in our sample always Small-Business (50 or fewer emp.), Mid-Market (51-1000 emp.) or Enterprise (> 1000 emp.).
date
When the review was posted. An ISO calendar date - 2025-03-16 - on all fifty-nine rows, never a relative phrase. Confirm it on your own run before you build a parser on it: four products is a narrow basis for a format claim.
helpful_count
Presumably how many people marked the review helpful. Never once populated in anything we hold.
verified
Whether the review is verified. Every row we have says Yes - we have never seen another value, so we cannot tell you what an unverified review looks like here.
review_link
Presumably a permalink to the individual review.
position
The order the review came back in, restarting at 1 for each run. The only way to preserve G2's ordering once the file is sorted.

Where these numbers come from. We hold twenty-seven run exports for this service. Eight returned data and nineteen came back empty, and the eight populated ones between them contain fifty-nine rows covering just four distinct products. So every measured statement above rests on fifty-nine reviews of four products, which is why the sample is repeated rather than tucked into a footnote. The column names are on firmer ground: the header row is identical across all twenty-seven workbooks, including the empty ones. The finding worth acting on is that three of the sixteen columns - reviewer_title, helpful_count and review_link - were empty on every single row. They are documented because they are in the schema, not because we have watched them work. Two more things we will not overstate: verified was Yes on all fifty-nine, so we cannot describe the alternative; and rating only ever came back '5.0' or '4.5', which tells you as much about the default sort as about the product. Run the free tier with the sort set to Lowest rated and you will learn more in one go than this page can honestly tell you.

Run one product free
Common workflows

Three jobs people
run this for.

All of them start from a product slug you can read off the URL bar.

Competitive

Read a competitor's cons column, and nothing else

Sort by Lowest rated, export, and read cons straight down. You get the specific things buyers dislike about a rival product, in their words, already separated from the praise - which is a more honest feature-gap analysis than any comparison page either vendor would publish.

Product · Strategy
Segmentation

Find out whether the problem is size-specific

Group cons by company_size and the picture usually splits. Complaints that dominate the Small-Business band are often onboarding and price; the Enterprise band complains about permissions, audit and scale. Knowing which you are hearing changes what you build.

Product · Research
Positioning

Borrow the language buyers actually use

pros is a column of people describing, unprompted, what a product is good for. Reviews here run to paragraphs rather than sentences, so there is enough text to mine for the phrases your own positioning should probably be using instead of the ones written in a meeting.

Marketing
Get started
Pricing

Pay per review row,
nothing else.

No subscription, no minimum, no per-seat licence. Your first 500 rows are on us - after that it is pay-as-you-go.

Free tier

500 free rows - $0

For every new account, one time. No credit card. All scrapers unlocked. A single product's review set usually sits well inside that, so the free tier covers a real first run against a competitor you actually care about.

$0 forever
Pay-as-you-go

$0.002 per row, after the free tier

Roughly $2 per 1,000 reviews, the same flat rate as every other scraper on the platform. The limit is per query, and leaving it empty fetches everything - worth setting a number while you are still exploring a large category.

Most popular
Enterprise

Custom - a whole category, on a schedule

Volume pricing, SLAs, dedicated workers and tailored onboarding for teams tracking an entire G2 category rather than a handful of competitors. Tell us your numbers and we will quote.

Talk to us
Start free See full pricing
10% off your first paid run.Use code LIVESCRAPER10 at checkout.
Sign up
Pairs well with

The same vendor,
judged elsewhere.

One review site is one audience. These are the others worth having beside it.

Product Hunt Reviews Scraper

Early adopters rather than procurement committees. The same software often reads very differently in the two places, and the gap between them is usually where the product outgrew its first audience.

Explore

Trustpilot Reviews Scraper

Reviews of the company rather than the software - billing, support, renewals. A vendor can score well on the product and badly on being a vendor, and buyers notice both.

Explore

LinkedIn Companies Scraper

The vendor behind the reviews - headcount, location, industry. Useful context when you are judging whether a competitor is scaling into your segment or retreating from it.

Explore
The legal bit

Is it legal to scrape
G2 reviews?

Short answer: yes, when you only collect what is publicly visible on a product page - and that is all this service reads.

A G2 product's review section is public. Anyone can read it without an account, and collecting publicly visible information for product and competitive 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.

These reviews describe people's working lives, though, and that makes the personal-data question sharper here than on a consumer site. reviewer_name is a person; company_size and the review text together can narrow down who they are and where they work. Publicly visible is not the same as free of obligation - if you store it, the GDPR and similar regimes apply regardless of where you got it. Analysing complaints in aggregate to improve a product is an easy case; building a file that identifies individual reviewers at named employers is not, and this service is not intended for it.

G2'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.

Sign up now
livescraper.app · principles
Public product pages only✓
No logins, no paywalls✓
Reviewer names and employers are personal data - aggregate, do not profile!
GDPR-aligned by default✓
Exports self-delete (30 days)✓
The same reviews any visitor sees on the product page.
Common questions

Things people ask before signing up.

The questions we hear most. Something else? Talk to us - humans write the answers, not bots.

What columns will the export contain?+
Sixteen: query, product_name, product_url, review_title, rating, review_text, pros, cons, reviewer_name, reviewer_title, company_size, date, helpful_count, verified, review_link and position. That is the header row of every export we hold, identical across all twenty-seven.
Do I need the full URL, or is a slug enough?+
A slug is enough, and we can say that from evidence rather than from the form's placeholder: we hold runs for the same product submitted once as a bare slug and once as the full reviews URL, and both returned that product's reviews. You can also upload a CSV, XLSX, TXT or Parquet file.
Are pros and cons really separate columns?+
Yes, and it is the main reason to use this over a general review scraper. G2 asks reviewers what they liked and what they disliked as two distinct questions, so the export carries pros and cons as their own columns rather than one blended paragraph. Both were filled on every row we hold.
Which columns should I not rely on?+
Three of the sixteen were empty on every one of the fifty-nine rows we hold: reviewer_title, helpful_count and review_link. They are part of the declared schema, but we have never seen a value in any of them, so check your own export before building anything that depends on them.
What does the rating column look like?+
It arrives as a string rather than a number. In our sample every value was either '5.0' or '4.5' - but our runs used the default G2 Sort, so that says as much about the ordering as about the product. Set the sort to Lowest rated if you want to know what the bottom of the distribution looks like.
Is the date a real date or a relative phrase?+
In all fifty-nine rows we hold it was a real calendar date in ISO form, like 2025-03-16, never something like '2 years ago'. That said, our sample covers four products, so confirm it against your own first run before writing a parser that depends on it.
What does the limit field do?+
It caps how many reviews come back per query, and on this form leaving it empty fetches all of them - that is the field's own hint. The minimum is 1, and unlike some of the other scrapers here it does not accept 0.
What does it cost?+
The first 500 rows on a new account are free and one-time; after that it is $0.002 per row - about $2 per 1,000 - pay-as-you-go with no subscription. Credits do not expire and there is no monthly reset.

Sort by lowest rated
and read the cons.

Paste a competitor's slug, set the sort, and read the one column that tells you what buyers actually dislike. Your first 500 rows are free.

Start free See full pricing
Activates instantly · no card required

Export G2 reviews as rows

G2 is where business software gets reviewed by the people who had to live with the procurement decision, and its reviews are unusually structured: the site asks separately what a reviewer liked and what they disliked, and records the size of the company they work for. That structure is what makes it worth exporting rather than reading. Paste a G2 reviews URL or just the product slug and each review comes back as a row - the title, the rating, the full text, pros and cons as their own columns, the reviewer's name and company size, the date, and whether the review is verified. Sixteen columns, one row per review.

Two controls shape the run. The limit is per query, and on this form leaving it empty fetches everything - the field says so itself, and it will not accept 0. The sort offers five settings, the first being G2's own default ordering. Changing it matters more here than on most sites: in the runs behind this page the default ordering returned nothing below 4.5 stars, so if you take the default you will conclude the product is beloved. Lowest rated is the setting that answers the question you probably came with.

The column to build on is cons. Elsewhere you would have to infer criticism from a paragraph that also contains praise; here it is already isolated, one row per reviewer, ready to group by company_size - which in our sample was always one of three bands, from small business through mid-market to enterprise. Complaints that cluster in one band are a different problem from complaints that appear across all three, and that distinction is usually the whole finding.

Two limits are stated plainly because they change what you should expect. First, three of the sixteen columns were empty on every row we hold - reviewer_title, helpful_count and review_link. They are in the schema; we have simply never seen them carry a value. Second, everything measured here comes from a narrow base: we hold twenty-seven run exports, of which eight returned data, and those eight contain fifty-nine rows spanning only four products. On that sample rating arrived as a string and never below 4.5, date was always an ISO calendar date, and verified was always Yes - real observations, but not a broad test, and this page would rather say so than round them up into promises. Run the free tier against a product you actually care about, with the sort set to Lowest rated, and you will learn more in one run than this page can honestly tell you. Your first 500 rows cost nothing and need no credit card. See pricing for current rates.