Trustpilot Reviews Scraper — Reviews, Ratings & Company Data

Scrape public Trustpilot reviews and company profiles with no login. $0.50 per 1,000 reviews on the Free plan (down to $0.30 on Gold), no start fee.

Trustpilot is protected by AWS WAF. The scraper solves the check once per proxy session in a headless browser, then reads pages over fast plain HTTP, so runs are quick and cheap (about 300 reviews per minute).

• Reviews: star rating, title, text, language, published / experience / updated dates, verification label, likes, reviewer (name, country, review count, avatar) and the company's reply.
• Companies: TrustScore, stars, total reviews, rating distribution 1–5, categories, website, email, phone, address, claimed / verified status, logo, and reply behaviour (reply rate, average days to reply).
• Filters: star ratings, languages, date range, verified only, with company reply only, sort (most recent / most relevant), and "reviews since".
• Discovery: search Trustpilot for companies, or list companies from category pages (country, minimum TrustScore and verified filters).
• Past the ~200-review limit: when you ask for more reviews than Trustpilot pages without login, the scraper splits the query by star rating and then by language (see Limitations).

Input examples

A single company (the prefill)

{ "startUrls": ["https://www.trustpilot.com/review/www.apple.com"], "maxReviewsPerCompany": 20 }

Plain domains and any country site

{ "startUrls": ["amazon.com", "https://uk.trustpilot.com/review/www.argos.co.uk", "de.trustpilot.com/review/zalando.de"], "maxReviewsPerCompany": 500 }

Only 1-star reviews from the last 3 months, in English

{
  "startUrls": ["www.example-airline.com"],
  "stars": ["1"],
  "languages": ["en"],
  "dateRange": "last3months",
  "maxReviewsPerCompany": 1000
}

Search companies, then 50 reviews from each of the top 5

{ "searchQueries": ["vpn", "web hosting"], "maxCompaniesPerSearch": 5, "maxReviewsPerCompany": 50 }

Category lead list (company items only, no reviews)

{
  "categoryUrls": ["https://www.trustpilot.com/categories/electronics_technology"],
  "categoryCountry": "US",
  "categoryMinRating": "4",
  "maxCompaniesPerSearch": 100,
  "includeReviews": false
}

Monitoring: new reviews since a date

{ "startUrls": ["www.example-shop.com"], "sortBy": "recency", "reviewsSince": "2026-10-01", "maxReviewsPerCompany": 5000 }

Output samples

Values below are invented.

Review

{
  "type": "review",
  "reviewId": "6f1c2a9e8b7d4c3a2b1e0f9d",
  "url": "https://www.trustpilot.com/reviews/6f1c2a9e8b7d4c3a2b1e0f9d",
  "companyName": "Example Electronics",
  "companyDomain": "www.example-electronics.com",
  "companyUrl": "https://www.trustpilot.com/review/www.example-electronics.com",
  "rating": 1,
  "title": "Order arrived two weeks late",
  "text": "The order arrived late and nobody answered my emails.",
  "language": "en",
  "publishedAt": "2026-10-27T09:14:02.000Z",
  "experiencedAt": "2026-10-24",
  "updatedAt": null,
  "isVerified": true,
  "verificationSource": "invitation",
  "verificationLevel": "invited",
  "isPending": false,
  "isEdited": false,
  "likes": 2,
  "reviewerName": "Jordan P.",
  "reviewerId": "8371b3f6e3308e4460413db2",
  "reviewerCountry": "GB",
  "reviewerReviewCount": 7,
  "reviewerImageUrl": null,
  "reply": { "text": "Sorry about the delay, Jordan. We've sent you a voucher.", "publishedAt": "2026-10-27T15:40:00.000Z", "updatedAt": null },
  "source": "www.example-electronics.com",
  "scrapedAt": "2026-10-28T11:25:26.339Z"
}

Company

{
  "type": "company",
  "businessUnitId": "4b70b35164d5581b2fa5eb4a",
  "name": "Example Electronics",
  "domain": "www.example-electronics.com",
  "url": "https://www.trustpilot.com/review/www.example-electronics.com",
  "trustScore": 4.3,
  "stars": 4.5,
  "totalReviews": 400,
  "ratingDistribution": { "1": 40, "2": 30, "3": 50, "4": 80, "5": 200 },
  "categories": ["Electronics Store", "Computer Store"],
  "primaryCategory": "Electronics Store",
  "description": null,
  "website": "https://www.example-electronics.com",
  "email": "support@example-electronics.com",
  "phone": "+1 555 0100",
  "address": { "street": "12 Example Street", "city": "Springfield", "zip": "99999", "country": "US" },
  "isClaimed": true,
  "isClosed": false,
  "verifiedBusiness": true,
  "logoUrl": "https://s3-eu-west-1.amazonaws.com/tpd/logos/4b70b35164d5581b2fa5eb4a/0x0.png",
  "replyBehavior": { "replyPercentage": 91.7, "averageDaysToReply": 1.4, "lastReplyToNegativeReview": "2026-10-20T08:15:00.000Z", "negativeReviewsWithRepliesCount": 88, "totalNegativeReviewsCount": 96 },
  "dataSource": "review_page",
  "source": "www.example-electronics.com"
}

Status (free)

{ "type": "status", "input": "missing-shop.com", "company": "missing-shop.com", "status": "not_found", "message": "No Trustpilot profile for \"missing-shop.com\"." }

status is one of:

• not_found: there is no profile.
• no_reviews: the company exists but has no reviews matching your filters.
• blocked: requests were still refused (403/429/challenge) after 4 retries.
• error: anything else.

Each run also stores a RUN_SUMMARY record. Per input it gives reviews saved, pages, splits used, blocks, proxy escalations and status, and per company each slice with its pages and why it stopped. It also gives totals, the percentage of each field that came back empty, and an overall status of ok, partial, no_results or failed.

Use cases

• Reputation monitoring: schedule a daily run with reviewsSince and get only the new reviews.
• Competitor analysis: compare TrustScore, rating distribution, reply rate and complaint themes across competitors.
• Sentiment and AI datasets: collect clean, deduplicated review text with ratings, languages and dates.
• Customer service QA: export negative reviews with and without replies (stars: ["1","2"], withRepliesOnly).
• Lead generation: list companies by category and country with website, email, phone and claimed status.

How it works and limitations

• Page cap and splitting. Without login, Trustpilot serves only 10 pages × 20 reviews (200 reviews) per filter combination; page 11 redirects to the login page.
  • When a company needs more than that, the scraper runs the same query once per star rating (1★ … 5★). Star buckets with 0 reviews are skipped when the rating distribution is known.
  • A star bucket that is still over the cap is split by language, using the languages Trustpilot lists for the company or a default list of 12 common languages.
  • Overlapping results are removed by reviewId.
  • The most reachable is about 200 reviews per star per language. For a very large company (for example 100k 5★ English reviews) you get the newest ~200 of each slice, not the full history.
  • With reviewsSince and "Most recent", the scraper only splits if the cap is actually hit before that date.
• Country sites. uk., de., fr. and other country URLs are read through www.trustpilot.com. The company data is the same; use languages to choose review languages.
• maxCompaniesPerSearch also applies to category pages.
• Listing data. With includeReviews: false, companies from search and category pages come from the listing data (dataSource: "listing"). That data has no rating distribution, claimed status or reply behaviour. Company URLs always use the full profile page.
• Dates. experiencedAt is a date (YYYY-MM-DD) and the other dates are ISO timestamps. "Reviews since" uses the published date.
• Only public reviews are collected. Reviews Trustpilot hides, removes or holds as pending don't appear.
• Proxy. Trustpilot blocks datacenter IPs outright (HTTP 403), so RESIDENTIAL proxy is the default and is required. On a 403/429 the scraper gets a new proxy session and a fresh WAF token, with backoff, up to 4 retries.
• Not found. An unknown company returns a free status item with not_found, confirmed once on a new session.
• Category filters. Trustpilot accepts trustscore=4.5 and 3.0/4.0 style values; "claimed profiles only" uses Trustpilot's claimed filter (there is no "verified company" filter on category pages).
• Replies filter. withRepliesOnly and verifiedOnly use Trustpilot's own replies=true / verified=true filters.
• Run timeout. The scraper stops 30 seconds before the run's timeout, saves what it has and exits successfully.
• Layout changes. Trustpilot changes its pages from time to time. All JSON paths and URL patterns are kept in src/mappings.ts.

Pricing

Pay per event. You only pay for results; there is no start fee and status items are free.

Event	Free	Bronze	Silver	Gold+
review, per 1,000 reviews	$0.50	$0.45	$0.38	$0.30
company, per 1,000 company items	$2.00	$1.80	$1.60	$1.40

Example: 1,000 reviews of one company cost about $0.50 on the Free plan. The run stops cleanly once your maximum charge is reached.

Development

npm install
npm test                  # unit tests on synthetic fixtures (+ live-capture tests when captured)
npm run mock-server       # offline Trustpilot mock on :4598 (see the header of test/mock-server.ts)
TP_BASE_URL=http://127.0.0.1:4598 npm start   # end-to-end against the mock, proxyConfiguration {"useApifyProxy": false}
npm run capture-fixtures  # save real Trustpilot responses to test/fixtures/live/ (needs network)
npm run fixtures          # regenerate the synthetic fixtures