Software Reviews Scraper - G2, Capterra, Trustpilot & TrustRadius

A Lantern Lab Actor. Name a software product once and get its public reviews from G2, Capterra, Trustpilot and TrustRadius in one run, all in one field layout: rating out of 5, title, review text, pros and cons, the reviewer's job title, industry and company size (never the name), publish date and link. Every product also gets one row per site with that site's rating, review count and what was collected. No login, no cookies of yours.

Price: you pay per review collected. Site rows, the run summary, a site that finds nothing and a site that fails are free. No start fee.

Quick example

Input: a product name finds it on G2 and TrustRadius; Trustpilot is found from the vendor's website; Capterra needs its link.

{
  "products": [
    { "name": "Slack", "capterraUrl": "https://www.capterra.com/p/135003/Slack/" },
    "Asana",
    "asana.com",
    { "name": "HubSpot CRM", "g2Url": "https://www.g2.com/products/hubspot-crm/reviews" }
  ],
  "maxReviewsPerSource": 50
}

One review row (recordType: "review", text shortened) from a real run:

{
  "recordType": "review",
  "source": "g2",
  "product": "Slack",
  "productName": "Slack",
  "sourceProductId": "3437",
  "reviewId": "10032434",
  "url": "https://www.g2.com/products/slack/reviews/slack-review-10032434",
  "rating": 5,
  "ratingOriginal": 5,
  "ratingScale": 5,
  "title": "Excellent",
  "body": "Slack is the best system out of all systems so far. It helps us make contact with our coll...",
  "pros": "Slack is the best system out of all systems so far. It helps us make contact with our coll...",
  "cons": "There is nothing that i do not like about Slack. So far I am personally happy with this sy...",
  "reviewerRole": "Payment Operations Specialist",
  "companySize": "Mid-Market (51-1000 emp.)",
  "publishedAt": "2024-08-21",
  "updatedAt": "2026-10-23",
  "partial": false,
  "partialReason": null,
  "collectedAt": "2026-10-24T02:39:58.270Z"
}

One site row per product (recordType: "source_status", free):

{
  "recordType": "source_status",
  "source": "trustpilot",
  "product": "Slack",
  "outcome": "found",
  "matchedBy": "vendor_homepage",
  "productName": "Slack",
  "productUrl": "https://www.trustpilot.com/review/slack.com",
  "siteRating": 2,
  "siteRatingScale": 5,
  "siteReviewCount": 360,
  "reviewsAvailable": 200,
  "reviewsRequested": 20,
  "reviewsReturned": 20,
  "partial": false,
  "partialReason": null
}

Limits

• Capterra needs capterraUrl. Capterra's site search is closed to crawlers by its robots.txt, so a name alone cannot find a Capterra product; that site row says needs_url. Copy the link from the address bar: https://www.capterra.com/p/<id>/<name>/.
• Trustpilot reviews a company, not a product. It is found by website domain: the domain you give, or the vendor website TrustRadius lists for the product (matchedBy: "vendor_homepage"). For a product of a large vendor that is the vendor's whole Trustpilot page. A vendor's Trustpilot page can exist with no reviews on it; that row is found with 0 reviews.
• Name matching is exact. G2 is asked for the product slug made from the name (Zoom Workplace -> zoom-workplace); TrustRadius for that slug, then its own search, taking only a product with exactly that name. When it is not found the row says so, with TrustRadius' closest names, and you can add the site link.
• What a visitor who is not signed in can see: G2 shows 100 reviews per product, Trustpilot 200, and Capterra shows some products' first 25 only. Asking for more marks that site row partial: true, partialReason: "sign_in_required"; nothing is reported as missing that the site does not show. TrustRadius has no such limit (up to 1,000 per product here).
• Review order: newest first on G2, Trustpilot and TrustRadius; Capterra's default order (most helpful), the only one it offers signed-out visitors.
• Public data only. Reviewer names, photos, profile links and company names are never collected. Trustpilot reviews carry the reviewer's country only.
• A run takes minutes: G2 pages take 20-50 seconds each through the proxy G2 requires, and Trustpilot and Capterra are read through a browser. Give it 2 GB of memory (the default).
• No scheduling or monitoring built in; use Apify schedules.

What it does

• Accepts per product a name, a website domain (a link to the vendor's own website counts as its domain), any of the four site links, or an object combining them. Several products per run (up to 100).
• Collects the four sites side by side. One site failing does not stop the others: its row is marked partial with the reason (blocked, upstream_error, ...), the other sites' reviews are written as usual.
• Writes every review in one layout. rating is always out of 5; TrustRadius rates out of 10, so its rating is halved and the original is in ratingOriginal with ratingScale: 10.
• Deduplicates within the run: a product reached twice is collected once, and a review is written and charged once.
• Says what happened on every site for every product: found, not_found, needs_url, invalid_url or partial, with how it was matched and why it is incomplete.
• Caps itself: maxUpstreamRequests is a hard ceiling on requests to all four sites together, derived from your input when left empty, so a run is never unbounded.

Input

Field	Default	Meaning
products	—	Product names, domains, site links, or objects {name, domain, g2Url, capterraUrl, trustpilotUrl, trustradiusUrl}.
sources	all four	Which of g2, capterra, trustpilot, trustradius to collect from.
maxReviewsPerSource	50	Reviews per product and site, max 1,000. 0 collects only the site rows.
maxUpstreamRequests	derived	Hard ceiling on requests to the four sites for the whole run, retries included.

Output

One row per review, recordType: "review", charged:

Field	Meaning
source	g2, capterra, trustpilot or trustradius.
product, productName, sourceProductId	The product as you named it, as the site names it, and the site's ID for it.
reviewId, url	The site's own review ID (the dedupe key, per site) and the review's link (Capterra: the product's reviews page).
rating, ratingOriginal, ratingScale	Rating out of 5, the site's own figure and its scale (5 or 10).
title, body, pros, cons	The review headline, its text and its pros / cons answers (G2: likes / dislikes). Trustpilot has no pros / cons.
reviewerRole, reviewerIndustry, companySize	The reviewer's profile where the site publishes it (G2, Capterra, TrustRadius). Never the name.
reviewerCountry, language	Trustpilot only.
vendorReply	The vendor's public reply (Capterra, Trustpilot).
publishedAt, updatedAt	When the site published and last updated the review, as the site states it (date or date-time).
partial, partialReason, collectedAt	Always false / null on a review row; our collection time.

One row per product and site, recordType: "source_status", free: outcome, matchedBy (url, name, domain, search, vendor_homepage), productName, productUrl, vendor, siteRating / siteRatingScale (G2, Capterra and Trustpilot's TrustScore out of 5, TrustRadius' score out of 10), siteReviewCount, reviewsAvailable (what a signed-out visitor can page through), reviewsRequested, reviewsReturned, duplicatesSkipped, partial / partialReason (upstream_request_cap, charge_limit_reached, blocked, upstream_error, sign_in_required, not_collected) and detail.

The last row is recordType: "RUN_SUMMARY": reviews per site, outcomes and partial reasons per site, requests per site, and why collection stopped. The same object is also in the run's key-value store under RUN_SUMMARY.