Software Reviews Scraper

Scrape B2B software reviews from Capterra, Software Advice, GetApp and TrustRadius and get every source back in the same fields. Pros and cons as separate columns, five sub-ratings, reviewer job title, industry and company size, vendor replies, incentive flags and the exact publication date. No login, no cookies, no API key.

What is Software Reviews Scraper?

Software Reviews Scraper is a software review API for four review sites: Capterra, Software Advice, GetApp and TrustRadius. Give it a product name or a product page, and it returns one row per review plus one summary row per product.

Most review scrapers cover a single site and invent their own field names, so comparing a product across two sites means writing a translation layer first. This Actor normalises everything: a review from Capterra and a review from Software Advice arrive with identical keys, so overallRating means the same thing wherever the row came from.

What you give it	What you get back
A product name, such as Slack	The matching product on each site you pick, with its reviews
A Capterra product URL	Every review, newest first, with pros, cons and sub-ratings
A Software Advice product URL	The same fields, same names
Several products at once	One dataset you can group by product and by site
A rating or date filter	Only the reviews you asked for

Which review sites are supported?

Site	Reviews available	Rating scale	Notes
Capterra	Full history, 25 per page	1-5	Richest source: sub-ratings, alternatives considered, switching reasons
Software Advice	Full history, 25 per page	1-5	Adds usage frequency and length of use
TrustRadius	Full history, 10 per page	1-10	Longest review text, plus reviewer company, department and years of experience
GetApp	The review showcase the page publishes (about 6)	1-5	Product ratings and totals are complete

Capterra, Software Advice and GetApp run on one review database, so a review carries the same reviewId wherever it appears — useful for de-duplicating across those three. TrustRadius is a separate site with its own reviewers.

Ratings on two different scales

TrustRadius rates out of 10 while the other three rate out of 5. Both numbers are returned: overallRating keeps the site's own scale, ratingScale says which scale that is, and normalizedRating converts everything to 1-5 so a sort or an average across sites means something. Rating filters work on the normalized value, so maxRating: 3 returns the same band of unhappy customers on every site.

What data can Software Reviews Scraper extract?

Field	Type	Example
source	string	capterra
sourceLabel	string	Capterra
productName	string	Slack
productUrl	string	https://www.capterra.com/p/135003/Slack/reviews/
productRating	number	4.4642
productReviewCount	integer	11145
reviewId	string	Capterra___7189737
title	string	Slack: Essential for Desk Based Workforces
publishedAt	string	2026-07-25
overallRating	number	5
easeOfUseRating	number	5
functionalityRating	number	5
valueForMoneyRating	number	4
customerSupportRating	number	5
recommendationScore	number	10
pros	string	Slack is the best way for a desk based workforce to communicate…
cons	string	Occasionally there will be a bug…
comments	string	I love Slack and could not run my business without it…
adviceToOthers	string	Start with fewer channels than you think you need.
reviewerName	string	Richard M.
reviewerJobTitle	string	Founder
reviewerIndustry	string	Marketing and Advertising
reviewerCompanySize	string	Self-employed
reviewerCompanySizeCode	string	C
reviewerCompany	string	SDLC Corp
reviewerDepartment	string	Information Technology
reviewerYearsExperience	integer	4
ratingScale	integer	10
normalizedRating	number	4.5
reviewUrl	string	https://www.trustradius.com/reviews/slack-2026-08-05-00-29-43
reviewerVerified	boolean	true
timeUsed	string	Less than 2 years
usageFrequency	string	Daily
incentivized	boolean	false
incentiveType	string	NominalGift
vendorResponse	string	Thanks for the detailed feedback…
reasonsForChoosing	string	Price and integrations
alternativeProducts	array	["Microsoft Teams", "Discord"]
switchedFrom	array	["Skype"]
ratingBreakdown	object	{"5": 17379, "4": 5725, "3": 911} (product rows)
status / errorCode / errorMessage	string	ok, or why a URL failed
scrapedAt	string	2026-09-05T04:20:11+00:00

A note on company size

Capterra publishes a readable range such as 11-50 employees, which lands in reviewerCompanySize. Software Advice and GetApp publish a single letter (A to I) instead, and neither site publishes what those letters mean. Rather than guess, the Actor passes the letter through in reviewerCompanySizeCode and leaves reviewerCompanySize empty. Nothing is invented.

How much does it cost to scrape software reviews?

Pay-per-event pricing, so the bill follows the rows you keep.

Event	Price	When it is charged
Review scraped	$0.0012	Per review written to the dataset
Product scraped	$0.003	Per product summary row

Failed URLs are never charged. A dead link, a page that does not exist, or a site that refuses the request comes back as an error row at no cost.

Job	Cost
1,000 reviews for one product	$1.20
10,000 reviews across 10 products	$12.03
200 recent reviews for 25 competitors	$6.08
Product ratings only, 100 products	$0.30

Comparable single-site review Actors charge between $0.90 and $5.49 per 1,000 reviews. This one sits near the bottom of that range while covering three sites in one schema. There are no proxy costs: pages are fetched over plain HTTPS.

The Apify Free plan includes $5 of monthly usage, which is about 4,000 reviews before you pay anything.

How to scrape Capterra reviews

1. Click Try for free and sign in to Apify.
2. Paste one or more product URLs into Product review URLs. A plain product link works; the reviews path is added for you.
3. Set Max reviews per product. Reviews come newest first, so 200 gives you the most recent 200.
4. Optionally set a rating or date filter — maxRating: 3 collects complaints only.
5. Press Start, then open the Output tab or export as JSON, CSV or Excel.

⬇️ Input

{
  "searchTerms": ["Slack", "Asana"],
  "searchSites": ["capterra", "trustradius"],
  "maxReviews": 200,
  "includeProductRow": true
}

Searching by product name

Put product names in 🔍 Product names and skip the URL hunt. Each name is looked up on every site in Sites to search, and the best match is scraped — so Slack returns the Slack reviews from Capterra and from TrustRadius in one dataset, ready to compare.

Matching prefers products whose name carries the words you typed, then the one with the most reviews, which keeps integrations and add-ons from stealing the slot: Zoom resolves to Zoom Workplace and its 14,000 reviews, not to a small app called ZooM. Every row that came from a search carries the searchTerm it was found by.

Raise Products per name above 1 to pull the closest alternatives alongside the product itself — a quick way to build a category comparison from a single keyword.

GetApp has no search page of its own and its URLs need a category that no other site publishes, so GetApp is reachable by URL only. A name that matches nothing on a site comes back as an error row naming the site and the term; the other sites still run.

Choosing products

Any product page on the four sites works. These are all valid:

• https://www.capterra.com/p/135003/Slack/
• https://www.capterra.com/p/135003/Slack/reviews/
• https://www.softwareadvice.com/project-management/slack-profile/reviews/
• https://www.getapp.com/collaboration-software/a/slack/reviews/
• https://www.trustradius.com/products/slack/reviews

Sorting what comes back

Sort reviews by orders the rows: newest, oldest, highest or lowest rating, or the order the site itself published them. It sorts what the run collected — the sites offer no sort of their own — so to control which reviews are collected, use the rating and date filters below. Asking for maxRating: 2 finds every one- and two-star review by paging until it has them, which is what a "lowest rating first" sort is usually a proxy for.

Collecting only the reviews you need

minRating and maxRating filter on the overall star rating. publishedAfter and publishedBefore take YYYY-MM-DD dates. Filters are applied while paging, so a run set to 300 complaints keeps reading until it has found 300 or the site runs out of reviews.

Proxies

Not required. The Actor reaches all three sites directly, which is why the price carries no proxy surcharge. The proxy field exists only as a fallback if a site starts refusing your particular runs.

⬆️ Output

One row per review:

{
  "rowType": "review",
  "source": "capterra",
  "sourceLabel": "Capterra",
  "productName": "Slack",
  "reviewId": "Capterra___7189737",
  "title": "Slack: Essential for Desk Based Workforces",
  "publishedAt": "2026-07-25",
  "overallRating": 5,
  "easeOfUseRating": 5,
  "functionalityRating": 5,
  "valueForMoneyRating": 5,
  "customerSupportRating": 5,
  "recommendationScore": 10,
  "pros": "Slack is the best way for a desk based workforce to communicate.",
  "cons": "Occasionally there will be a bug or issue, but that is once in a blue moon.",
  "reviewerName": "Richard M.",
  "reviewerJobTitle": "Founder",
  "reviewerIndustry": "Marketing and Advertising",
  "reviewerCompanySize": "Self-employed",
  "reviewerVerified": true,
  "incentivized": false,
  "status": "ok",
  "scrapedAt": "2026-09-05T04:20:11+00:00"
}

And one summary row per product:

{
  "rowType": "product",
  "sourceLabel": "Software Advice",
  "productName": "Slack",
  "productRating": 4.6639,
  "productReviewCount": 24190,
  "ratingBreakdown": { "5": 17379, "4": 5725, "3": 911, "2": 127, "1": 48 },
  "ratings": { "overall": 4.6639, "easeOfUse": 4.6114, "customerSupport": 4.4286 }
}

Why one schema matters

Run the same product on two sites and the rows line up column for column. Capterra rates Slack 4.46 across 11,145 reviews; Software Advice rates the same product 4.66 across 24,190. That gap is visible in one sort, without reconciling two field naming schemes first.

Error rows

errorCode	Meaning
unsupported-site	The URL is not one of the three supported sites
not-found	The product page does not exist
blocked	The site refused the request; retry, or add a proxy
http-error	The site answered with an unexpected status
network-error	The site could not be reached

Use cases for software review data

Competitor and win-loss research

Pull every review of your competitors, filter to maxRating: 3, and read what their customers complain about. reasonsForSwitching and switchedFrom name the products buyers left behind.

Product feedback and roadmap input

Collect your own reviews with publishedAfter set to the last quarter and group cons by reviewerIndustry or job title to see which segment is unhappy.

Feeding an LLM or a RAG pipeline

Every review arrives as clean text with a date and a rating, which is what a summarisation or sentiment prompt needs. Because pros and cons are separate fields, you can prompt each side independently instead of asking a model to split the text again.

Lead generation and sales research

reviewerJobTitle, reviewerIndustry and the products a reviewer switched away from describe a buying decision already made in public.

Market and category monitoring

Schedule a run weekly across a category, keep the product rows, and track how average ratings and review volumes move over time.

Integrations

• API: POST https://api.apify.com/v2/acts/lergassy~software-reviews-scraper/runs
• Python: ApifyClient(token).actor('lergassy/software-reviews-scraper').call(run_input={'productUrls': [...]})
• JavaScript: await client.actor('lergassy/software-reviews-scraper').call({ productUrls: [...] })
• n8n: use the Apify node and set a max total charge so a scheduled run cannot exceed your budget.
• Make, Zapier, Google Sheets, Airtable: map the dataset straight into a sheet or table.
• MCP: available through the Apify MCP server, so Claude, Cursor and other MCP clients can call it as a tool.
• Webhooks: fire a webhook when a run finishes and pass the dataset onward.

🤖 For AI agents and LLM apps

{ "productUrls": ["https://www.capterra.com/p/135003/Slack/reviews/"], "maxReviews": 50 }

• One row per review; rowType separates review, product and error rows.
• Field names never change between sites, so a prompt written once works for all three.
• pros and cons are separate strings — no text splitting needed.
• incentivized flags reviews written in exchange for a gift, which matters when weighing sentiment.
• Failed URLs are returned as rows and are never charged, so agent retries stay cheap.

❓ FAQ

Is it legal to scrape software reviews?

The Actor reads pages that are public and require no login, and it collects the review text and the reviewer details each site already publishes. How you may store and use that data depends on your jurisdiction and on the sites' terms, so review both before running at scale. This is not legal advice.

Do I need a proxy or an API key?

No. The three sites are reached directly, with no key, cookie or proxy.

How many reviews can I get for one product?

As many as the site publishes. Capterra and Software Advice page through their whole history 25 reviews at a time; popular products have tens of thousands. GetApp publishes only a short showcase, and the Actor says so in the run log rather than pretending otherwise.

Can I search by product name instead of hunting for URLs?

Yes. Put the names in Product names and pick the sites in Sites to search. Capterra, Software Advice and TrustRadius can all be searched; GetApp needs a product URL because it publishes no search page. Searching costs nothing extra — you pay for the reviews and product rows you get back.

Does it scrape G2?

No. G2 sits behind a bot-protection service that this Actor deliberately does not try to defeat. Capterra, Software Advice, GetApp and TrustRadius cover the same buyers and are reachable cleanly.

How are TrustRadius scores out of 10 handled?

Both ways. overallRating keeps the 9-out-of-10 the site published, ratingScale records that it was a 10-point scale, and normalizedRating gives you 4.5 so the row lines up with a Capterra review.

Can I get only negative reviews?

Yes: set Maximum star rating to 2 or 3. The Actor keeps paging until it has collected the number you asked for.

Can I use it with the Apify API?

Yes, like any Apify Actor: start a run, then read the dataset.

Can I use it through an MCP server?

Yes, it is exposed through the Apify MCP server as a callable tool.

Why is company size sometimes a single letter?

Because that is what Software Advice and GetApp publish. See the note above — the letter is passed through untouched rather than guessed at.

How fresh is the data?

Reviews are read live at run time, newest first, so a run reflects the site at that moment.

Your feedback

Missing a field, or a site you need covered? Open an issue on the Issues tab or leave a review — issues are answered quickly, and requests shape what gets built next.

You might also like

Actor	What it does
Document Text Extractor	PDF, Word, Excel and PowerPoint to Markdown, tables and RAG chunks
Email & Phone Verifier	Validate e-mail addresses and phone numbers, find contacts on a website
US Business Leads	New business registrations from US state and city registries
Agoda Reviews Scraper	Hotel reviews and ratings from Agoda and Booking.com