Software Review Aggregator | G2, Capterra & TrustRadius in one schema

Collect B2B software reviews from G2, Capterra and TrustRadius and get them back in a single, normalized JSON schema: platform, product, rating (always on a 0-5 scale), review_title, body, date and reviewer. Use it for competitive research, voice-of-customer analysis, sales battlecards, product-market-fit checks, or to feed review text into an LLM pipeline.

The Actor is built for agents and pipelines: one structured call, predictable output, and you only pay for reviews that are actually returned.

What you get

For every review the dataset contains:

Field	Description
platform	g2, capterra or trustradius
product	Product name as shown on the review site
product_url	Listing URL the review was read from
rating	Rating normalized to 0-5 (rating_raw and rating_best keep the original scale)
review_title, body	Review headline and text (Capterra pages expose excerpts and no headline)
date	Publication date as published by the site
reviewer	Reviewer display name
source	live or wayback (see below)
snapshot_date	Date of the archived snapshot when source is wayback

How it works

1. Every input is resolved to a canonical reviews listing URL (productUrls[] or products[] with { platform, slug }).
2. Pages are fetched over plain HTTP with got-scraping (browser-like TLS and headers), optionally through your proxy. No headless browser is used.
3. Reviews are parsed from schema.org JSON-LD (G2, TrustRadius), microdata, or the Markdown rendition that Capterra serves to some clients.
4. Review platforms protect themselves with DataDome and Cloudflare. When a live request is blocked or returns no review markup, the Actor falls back to the closest Internet Archive (Wayback Machine) snapshot of the same page, so you still get usable data. Archived data is older and usually covers fewer pages, which is why every row carries source and snapshot_date.
5. Pagination (?page=N) continues until maxReviews, maxPagesPerProduct or the end of new reviews is reached. Duplicates are removed.

For the most complete and freshest results, run live with a residential proxy in proxyConfiguration.

Input

{
  "productUrls": ["https://www.g2.com/products/slack/reviews"],
  "products": [
    { "platform": "capterra", "slug": "135003/Slack" },
    { "platform": "trustradius", "slug": "slack" }
  ],
  "maxReviews": 50,
  "maxPagesPerProduct": 5,
  "since": "2026-01-01",
  "useWaybackFallback": true
}

• since skips older reviews, and skipped reviews are not charged.
• Set useWaybackFallback to false if you only want live data.
• Capterra slugs are <product id>/<Name>; TrustRadius and G2 slugs are the product slug.

Pricing (pay per event)

Event	Price
Actor start	$0.02
Review found	$0.002

Example: 500 reviews cost 500 x $0.002 + $0.02 = $1.02. Rows that report an error or a blocked page are free, and Apify's "max total charge" limit is respected: the run stops cleanly when the limit is reached.

Blocked or empty results

When a site cannot be read (no live access and no archive snapshot), you receive a free dataset row with error, blocked: true and a hint instead of a crash. The rest of your products still run. Typical fixes: add a residential proxy, retry later, or verify the product URL.

Notes and limits

• Only publicly visible review content is collected. Respect each site's terms of service and applicable data protection law when storing reviewer names.
• Review site layouts change. The parser relies on stable schema.org markup rather than CSS classes, which keeps breakage low, but no scraper is permanent.
• Capterra Markdown excerpts do not include per-review star ratings; rating is null for those rows.

Run locally

cd apify-software-review-aggregator
echo '{"productUrls":["https://www.g2.com/products/slack/reviews"],"maxReviews":10}' > storage/key_value_stores/default/INPUT.json
apify run