G2 & Capterra Software Reviews Scraper

Scrape software reviews from G2 and Capterra into one clean, normalized dataset. Paste product review URLs from either site (or mix both in one run) and get every review with the same fields: star rating, date, title, pros, cons, full text, reviewer role, company size, industry, and incentive / verification flags.

Built for SaaS marketers, product marketers, and competitive-intelligence teams who need review data they can drop straight into a spreadsheet, BI tool, or LLM pipeline without writing two different parsers.

What you get

• Both sites, one schema. G2 and Capterra reviews share the same field names, so you can compare your product against competitors across both sites in a single table. Fields a site does not publish are null, never guessed.
• Pros and cons kept separate (plus "problems solved" on G2 and the overall comment on Capterra), and a combined fullText field for sentiment analysis or LLM summarization.
• Reviewer context: job title / role, company size, industry, how long they used the product (Capterra), anonymous or named.
• Trust signals: isIncentivized, isVerified, verification methods, review source (e.g. G2 "Organic" or "Seller invite"), and G2's "current user" badge, so you can filter out incentivized reviews if you need to.
• Capterra extras: sub-ratings (ease of use, customer support, features, value for money), likelihood-to-recommend score, products the reviewer switched from and why, and vendor responses.
• Filters: max reviews per product, star ratings (e.g. only 1–2 star reviews), date range (since / until, absolute or relative like 90 days), and newest-first sorting on G2 (see Capterra note under Tips & limitations).
• Optional product summary rows: average rating, total review count, categories, and G2's top pros/cons themes when G2 includes them in the page it serves.
• Reliable by design: HTTP-only (no browser), Apify's unblocking proxy, automatic retries with fresh sessions, detection of half-loaded pages, and per-product error handling: one bad URL never kills the whole run.

Use cases

• Competitor teardown: pull the newest 1–2 star reviews of your top competitors and see what their customers complain about.
• Battlecards and positioning: collect pros/cons by company size or industry to sharpen messaging.
• Voice-of-customer research: feed fullText into an LLM to find recurring themes, objections, and phrases buyers actually use.
• Review monitoring: schedule a weekly run with "Only reviews since: 7 days" to track new reviews of your product and competitors.
• Ad and landing-page copy: mine real customer language from pros sections.

How to use

1. Add one or more product review URLs to Products, for example:

   • https://www.g2.com/products/slack/reviews
   • https://www.capterra.com/p/135003/Slack/

   Shortcuts also work: g2:slack and capterra:135003/Slack.

2. Set Max reviews per product (you pay per review saved, so start small).

3. Optionally pick star ratings, a date range, and the sort order.

4. Click Start. Download results as JSON, CSV, Excel, or HTML, or pull them via the Apify API or integrations (Make, Zapier, Google Sheets, webhooks, and more).

Example input

{
    "products": [
        "https://www.g2.com/products/slack/reviews",
        "https://www.capterra.com/p/135003/Slack/"
    ],
    "maxReviewsPerProduct": 100,
    "sort": "newest",
    "ratings": ["1", "2"],
    "reviewsSince": "180 days"
}

Example output (one review)

{
    "recordType": "review",
    "source": "g2",
    "productName": "Slack",
    "productId": "3437",
    "productUrl": "https://www.g2.com/products/slack/reviews",
    "reviewId": "13217252",
    "reviewUrl": "https://www.g2.com/products/slack/reviews/slack-review-13217252",
    "title": "Easy Team Connection: Presence, Huddles, and Effortless Groups",
    "rating": 5,
    "ratingScale": 5,
    "date": "2026-10-30",
    "reviewerRole": "Senior customer success executive",
    "reviewerCompanySize": "Mid-Market (51-1000 emp.)",
    "reviewerIndustry": null,
    "pros": "We can connect with anyone from the company ...",
    "cons": "The only thing I dislike is there is no option available to see whether a person has seen the message or not ...",
    "problemsSolved": "It solves basic problems like getting details of any unknown executive ...",
    "fullText": "What do you like best? ...",
    "isIncentivized": false,
    "isVerified": true,
    "reviewSource": "Organic",
    "scrapedAt": "2026-10-31T15:00:00.000Z"
}

The full field list with descriptions is in the Output / dataset schema tab. The dataset has ready-made views: Overview, Pros & cons, Reviewers, and Product summaries.

Output fields (summary)

Field	Description
source	g2 or capterra
productName, productId, productUrl	Product being reviewed
reviewId	Review identifier on the source site
reviewUrl	Direct link to the review (G2 only; Capterra has no per-review link, so the field is left out of Capterra rows)
pageUrl	The review-list page the review was read from (both sites)
rating, ratingScale	Star rating out of 5, exactly as the site displays it. G2 uses half stars from 0 to 5, so a very negative G2 review can genuinely be 0 or 0.5 (these fall in G2's own "1 star" filter); Capterra uses 1–5
date	Publish date (YYYY-MM-DD), both sites
dateTime	Full publish timestamp with time zone (G2 only; Capterra only publishes a date, so the field is left out of Capterra rows)
title, pros, cons, fullText	Review text (both sites)
problemsSolved, reviewerProfileUrl, isCurrentUser, reviewerNotice	G2 only
reviewerName, reviewerRole, reviewerCompanySize, reviewerIndustry	Reviewer context, as published by the site
isIncentivized, isVerified, verificationMethods, reviewSource	Trust signals
summaryComment, reviewerTimeUsedProduct, subRatings, recommendationRating, switchedFrom, reasonsForSwitching, vendorResponse	Capterra only

Fields that a site never provides are left out of that site's rows instead of being filled with empty values. Fields a site does provide can still be null when the reviewer skipped that question.

A RUN_SUMMARY record in the key-value store lists each product's status (ok, partial, failed), review count, pages read, and any error message.

Pricing

This Actor uses pay-per-event pricing: you pay a fixed price per review saved to the dataset (see the Pricing tab for the current price). Product summary rows are free. You are not charged for retries, blocked pages, or failed products.

To control spend, set Max reviews per product and, in the run options, Max cost per run. When the cost limit is reached the Actor stops cleanly and keeps everything saved so far.

Tips and limitations

• Use the default proxy setting (Apify UNBLOCKER group). G2 and Capterra both use strong bot protection; plain datacenter or residential proxies are usually blocked.
• G2 star filter is applied by G2 itself. G2 groups half stars upward (a 4.5-star review counts as 5 stars).
• Capterra sorting: Capterra is read in its default "Most Helpful" order; newest-first is not available there (we tried, and Capterra did not honor it). Rating and date filters are applied while paging, so narrow filters on Capterra can require reading many pages (capped by Max listing pages per product) and may return fewer reviews than requested.
• Capterra speed: Capterra pages load through the unblocker proxy in roughly 10–25 seconds each, and some requests stall. The Actor cuts a Capterra request off after 35 s (configurable), starts a parallel request on a fresh proxy session if the first has not answered after 15 s, and retries timeouts almost immediately. This shortens stalls but does not remove them: a Capterra product can still take a few minutes, and on a bad run a later page can fail, in which case the product is marked partial and the reviews already saved are kept.
• G2 pages sometimes come back without the review list; the Actor detects that and retries (also via G2's lighter review-list URL), which can add time but not charges.
• Capterra caps listings at 100 pages (25 reviews per page). G2 shows 10 reviews per page.
• Capterra has no per-review links or timestamps, so Capterra rows have no reviewUrl or dateTime; use pageUrl/productUrl plus reviewId, and date.
• No "helpful" vote counts: neither site exposes them in the pages this Actor reads, so there is no such field.
• Fields depend on what each reviewer filled in and what the site displays publicly; unanswered questions are null.
• Site layouts change. If a site changes its pages, the run will report the product as failed with a clear message rather than returning broken data.

FAQ

Is scraping G2 and Capterra legal? This Actor collects publicly visible review content only and does not log in. Reviews can contain personal data such as reviewer names. Make sure your use complies with applicable laws (e.g. GDPR) and the sites' terms. If unsure, consult a lawyer.

Can I scrape many products at once? Yes. Add as many URLs as you like; products run in parallel (configurable) and each one is handled independently.

Can I get only negative reviews? Yes: select 1 and 2 stars under Star ratings to include.

Something doesn't work? Open an issue on the Actor's Issues tab with the run ID and product URL, and we'll look into it.