G2 Review Scraper (Apify Actor)

Commercial G2 review scraper built with Crawlee (Node.js) + PlaywrightCrawler.

Core selling points:

• star_rating_max: precise inclusive filter (e.g. 2 keeps only 1–2 star reviews)
• sort_by: request-time sorting (e.g. newest)

Project structure (Apify Store standard)

• actor.json: Actor metadata
• Dockerfile: Apify runtime image + build/run commands
• .actor/input_schema.json: Store UI input schema
• src/main.ts: Actor entry (Crawlee + PlaywrightCrawler)
• src/routes/reviews.ts: G2 review extraction + filtering
• src/utils/g2.ts: URL normalization + parsers

Inputs

Example input:

{
  "startUrls": [{ "url": "https://www.g2.com/products/slack/reviews" }],
  "sort_by": "newest",
  "star_rating_max": 2,
  "maxReviews": 1000,
  "maxPagesPerUrl": 50,
  "proxyConfiguration": { "useApifyProxy": true }
}

Stealth mode + Apify Proxy (Cloudflare)

This Actor uses:

• Stealth: playwright-extra + playwright-extra-plugin-stealth
• Apify Proxy: Actor.createProxyConfiguration(input.proxyConfiguration)
• Session pool: enabled (useSessionPool, persistCookiesPerSession) so failed sessions get rotated

Recommended for Cloudflare:

• Use Apify Proxy Residential (and optionally countryCode)
• Keep concurrency low (default is maxConcurrency: 3)

Star rating & date parsing

• Star rating is extracted from aria-label / text patterns such as:
  • "4.0 out of 5", "4 stars", "Rated 4 out of 5"
• Date is extracted from <time datetime="..."> when available and stored as ISO string.

Filtering rule:

• Reviews are kept only when starRating <= star_rating_max
• If star rating is missing, the review is dropped (commercial correctness over recall).

Notes on sort_by

G2 may change query parameter names/values. Sorting is centralized in:

• src/utils/g2.ts → normalizeG2ReviewsUrl()

If you notice sorting isn’t applied, adjust the mapping there to match current G2 behavior.