Trustpilot Review Scraper

Pull every public review from any Trustpilot company page — past the 200-result UI cap — straight into CSV or JSON. No API waitlist, no auth, no headless browser.

949 lifetime production runs on this exact actor as of 2026-04-29.

---

What you get per review

Every record returned is a flat JSON object with these 9 fields (verified against src/main.js Actor.pushData()):

Field	Example
companyName	Booking.com
companyUrl	https://www.trustpilot.com/review/booking.com
title	Review headline (may be empty string if missing in JSON-LD)
text	Full review body, plain text (may be empty string if missing)
stars	1 – 5 integer, or null when reviewRating.ratingValue is non-numeric
author	Reviewer display name (may be empty string)
datePublished	ISO 8601 (2026-03-12T08:14:00Z) — or null if absent
inLanguage	en-US, de-DE, etc. — or null
scrapedAt	ISO 8601 capture timestamp

CSV output is the same fields as a header row. No nested arrays — opens cleanly in Excel, Google Sheets, pandas, dbt seeds.

---

Use it for

• Competitor analysis — pull a rival's 1-star reviews and tag them by complaint category
• Sentiment monitoring — push the text column into your LLM of choice; track shifts week-over-week
• Lead enrichment — companies with rising 1-star volume often hire support / CX consultants
• Market research — feed inLanguage + stars into a regional satisfaction dashboard
• Reputation early-warning — schedule a daily run, diff the result, alert on new negative reviews

---

Quick start

1. Click Try for free above.
2. Paste a Trustpilot company URL into companyUrls (e.g. https://www.trustpilot.com/review/booking.com).
3. Optionally set maxReviewsPerCompany (default 100, max 1000 per input_schema).
4. Run. Download CSV/JSON from Storage → Dataset.

For programmatic use (input keys match .actor/input_schema.json exactly):

from apify_client import ApifyClient

client = ApifyClient("YOUR_APIFY_TOKEN")
run = client.actor("knotless_cadence/trustpilot-review-scraper").call(
    run_input={
        "companyUrls": ["https://www.trustpilot.com/review/booking.com"],
        "maxReviewsPerCompany": 1000,
        "filterByStars": 0,
        "sortBy": "recency",
    }
)
for review in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(review["stars"], review["title"])

Inputs:

• companyUrls — array of Trustpilot review URLs (or bare domains; the actor normalises to https://www.trustpilot.com/review/<domain>)
• searchQueries — array of search terms (auto-discovers companies + scrapes top 10 per query — hard .slice(0, 10) cap)
• maxReviewsPerCompany — integer cap per company URL (default 100, schema max 1000). Multi-URL runs multiply: 5 URLs × 100 = up to 500 reviews.
• filterByStars — 0 = all, 1–5 = only that rating. Records outside the filter are dropped entirely (not emitted with partial schema).
• sortBy — recency (default) or relevance. The actor appends ?sort=${sortBy} directly to Trustpilot URLs.

---

How the 200-cap bypass works

Trustpilot's public UI paginates reviews; older entries are not always reachable from the default sort view. This actor walks ?page=N directly using CheerioCrawler (HTTP only — no headless browser, no captcha layer), parses the JSON-LD Review graph that Trustpilot embeds in every page, and pushes each review as a flat row. Concurrency is capped at maxConcurrency: 5 and per-crawl requests at maxRequestsPerCrawl: 1000.

---

Honest limitations (verified against src/main.js)

• No Apify Proxy is wired. CheerioCrawler is constructed without proxyConfiguration. Requests originate from Apify's default IP pool and can be rate-limited at high concurrency. For proxy-routed runs at scale, request a custom build via the contact below.
• No headless rendering. Currently Trustpilot embeds the Review JSON-LD graph server-side — Cheerio reads it directly. If Trustpilot moves to client-only rendering, this actor will return empty results. Disclosure-honest: this is a fragility, not a guarantee.
• sortBy is appended as ?sort=recency / ?sort=relevance. Trustpilot's UI may use different sort tokens internally; verify your sort actually changes ordering before relying on it for time-series work.
• Pagination terminates on the first empty page. Transient server [] (network blip, brief block) on page N silently ends pagination for that company. No retry on the pagination boundary; missing tail-pages are not re-attempted within the same run.
• Crawlee internal retries: 3 per request for transient failures (default). Persistent blocks return a silent zero-record run, not an exception.
• Search dedup is exact-URL match. Case differences or trailing-slash variants in href attributes are treated as different companies; you may see one extra duplicate per search.
• maxReviewsPerCompany is per-URL. Whole-run total = len(companyUrls) × maxReviewsPerCompany plus up to len(searchQueries) × 10 × maxReviewsPerCompany from auto-discovered companies. Budget accordingly.
• filterByStars halves the dataset effective yield when you set a single rating — pagination still consumes the full maxReviewsPerCompany budget even though most rows are dropped. To guarantee N records of a specific rating, raise the cap.
• Schema fields can be empty strings or null (see table above). Downstream code should handle text == '' and stars is None rather than asserting populated values.

---

Need something custom?

If your use case is broader than "give me Trustpilot reviews" — multi-source review aggregation, scheduled deltas piped to your DB, data cleaned and joined to your CRM — we build that as a one-shot package.

Tier	Price	Includes
Pilot	$97	1 custom actor or modification, 7-day support
Standard	$297	Custom actor + Slack/email alerts on results, 30-day support
Premium	$797	Custom actor + dashboard + 90-day support + 1 modification round

Email: spinov001@gmail.com Blog (case studies + writeups): https://blog.spinov.online Telegram channel (scraping & data engineering tips): https://t.me/scraping_ai

Pilot pricing while we grow our public portfolio. Most pilots delivered inside 48–72 hours.

---

Honest disclosure

• Public reviews only. We do not scrape private user data, accounts, or anything behind a login.
• Independent project — not affiliated with Trustpilot.
• This actor is maintained by the same author who runs apify.com/knotless_cadence (78 actors, 31 public).