Unified B2B Reviews Scraper (Trustpilot)

Company reviews from Trustpilot in one clean schema — for sentiment analysis, reputation monitoring and B2B lead generation.

No logins. No browsers. No bloated output. Just normalized, GDPR-minimized review data, ready for your pipeline.

Why this actor

• One unified schema — every review comes out normalized (rating, text, dates, country, verified, company trust score). Built for G2/Capterra expansion: same schema, more platforms.
• GDPR-minimized by design — author names, IDs and avatars are never collected. Only rating, text, dates, country and review counts. Safer for EU-facing products.
• Fast and cheap — pure HTTP + server-rendered data extraction (no headless browser per page). Runs through Apify Unblocker to pass anti-bot walls reliably.
• Full review metadata — verification level, incentivized/paid-review flag, organic vs invited, company reply flag, experienced vs published date, consumer country and review history.
• Empty result = free — you pay only for reviews actually delivered.

Input

Field	Type	Default	Description
startUrls	list	—	Trustpilot review URLs, e.g. https://www.trustpilot.com/review/booking.com
domains	list	—	Just domains: booking.com (alternative to startUrls)
maxReviews	int	1000	Max reviews per domain
maxPages	int	100	Max pagination depth per domain
reviewsSince	string	—	Keep only reviews newer than this: ISO date (2026-10-01) or relative (7d, 30d, 6m, 1y)
minStars / maxStars	int	0	Rating filter (1–5). minStars: 1, maxStars: 1 → only 1-star complaints
languages	string[]	[]	Keep only reviews in these language codes, e.g. ["en", "ru"] (case-insensitive, en also matches en-GB). Empty = all languages.
sort	string	—	latest = newest first (pairs well with reviewsSince)
includeRaw	bool	false	Add _raw: original review JSON from Trustpilot (custom fields, debugging)
proxy	proxy	Apify Proxy	Residential/Unblocker recommended — Trustpilot blocks datacenter IPs

Sample output (dataset)

{
  "platform": "trustpilot",
  "reviewId": "6ad7141972d884acd5c59149",
  "companyName": "Booking.com",
  "companyDomain": "www.booking.com",
  "companyTrustScore": 1.4,
  "companyReviewCount": 129045,
  "rating": 1,
  "title": "If I could only give -100 stars to…",
  "text": "Full review text…",
  "language": "en",
  "likes": 0,
  "source": "Organic",
  "verified": false,
  "verificationLevel": "not-verified",
  "incentivized": true,
  "publishedDate": "2026-10-18T20:40:56.000Z",
  "experiencedDate": "2026-10-18T00:00:00.000Z",
  "consumerCountry": "US",
  "consumerReviewsCount": 19,
  "hasCompanyReply": true,
  "companyReply": { "text": "Hi there, we really appreciate…", "publishedAt": "2026-10-19T18:59:27.000Z" },
  "url": "https://www.trustpilot.com/review/booking.com?reviewId=…"
}

Named output (OUTPUT): { "totalReviews": "80", "domains": "[{\"domain\":\"booking.com\",\"reviews\":40,\"pagesFetched\":2,\"ok\":true}]" }

Use cases

• Competitive intelligence — track review velocity and sentiment for rival products
• Reputation monitoring — scheduled runs + webhooks = alerts on new reviews
• B2B leadgen — find companies whose customers complain about specific pain points (that your product solves)
• LLM/RAG pipelines — clean schema drops straight into embeddings, no cleaning step

Notes

• Trustpilot pages hold ~20–25 reviews per page; the actor paginates with per-page deduplication.
• Recommended proxy groups: RESIDENTIAL or UNBLOCKER (charged to your account per Apify pricing).
• Fair use: respect Trustpilot's terms, scrape public data only.

Roadmap

• G2 Reviews (same schema)
• Capterra Reviews (same schema)

Feedback and feature requests → the Support tab. Built and maintained with automated daily quality canaries: every listing claim is verified against live runs.