Trustpilot Reviews Scraper - Brand Ratings & Replies

Turn public Trustpilot reviews into a brand research table for US CX teams, ecommerce agencies and support analysts. Export ratings, review text, company replies and observed reply timing; retrieve overall TrustScore summaries separately. No Trustpilot login or API key required.

Use company domains or review slugs such as nike.com, or full Trustpilot company review URLs. The dataset contains public review rows, while the current run's COMPANIES key contains company-level metrics. These are different sources for your reporting workflow, not a ready-made sentiment dashboard.

Current pricing: $1 per 1,000 successfully saved reviews + $0.00005 per Actor Start event. The already scheduled start fee becomes $0.00030 on 23 October 2026 at 21:20 UTC (24 October at 02:50 IST); review pricing and included platform usage stay unchanged. Apify support confirmed cancellation of the former $0.04/company-check proposal, and the company-check event has been removed from the manifest. The live Apify Pricing tab remains authoritative; this documentation package does not change pricing.

The Actor uses a browser because Trustpilot can present anti-bot challenges. Direct access is the default to control cost; proxy settings remain available when a user explicitly needs them. The Actor never silently switches a run to Residential traffic. It saves reviews to the default dataset and parsed company summaries to the current run's COMPANIES key.

US-first describes the buyer workflow, not US-only reviewers or verified US customers. Reviews on a company page may be global. The Actor has no reviewer-country filter, built-in sentiment scoring, automatic alerts, or generated CX report, and does not provide persistent watchlist history. Repeated-run comparisons and reports happen in your own spreadsheet or workflow system.

US brand review research

Use the US Brand Review & Reply Research task to start with two illustrative brands, Nike and Adidas. Replace these domains with appropriate company review pages for your research; example inclusion is not an endorsement or a claim that the brands' review populations are comparable.

The starter requests up to ten recent reviews per brand, all ratings, without a verified-only filter or proxy. It uses the existing 1024 MB browser configuration. A source block, spending limit, unavailable page or exhausted matching feed can return fewer than twenty rows; successful status alone does not establish complete coverage.

{
  "companyNames": ["nike.com", "adidas.com"],
  "companyUrls": [],
  "maxReviewsPerCompany": 10,
  "sortBy": "most_recent",
  "filterByRating": "all",
  "verifiedOnly": false,
  "proxyConfiguration": { "useApifyProxy": false }
}

If twenty reviews are successfully saved, listed Actor events total $0.02005 now, or $0.02030 after the already scheduled start-fee change, at 1024 MB. Fewer saved reviews reduce the review-event charge. Company summaries have no separate company-check charge. These are fee calculations, not measured operating costs or a profitability guarantee; any explicit proxy use and account subscription have separate implications.

This is a saved configuration, not newly live tested for this documentation package. It creates no schedule, stored historical baseline, comparison report or automatic customer outreach. The existing Nike task remains unchanged. Start with the one-review quick start below if you first want a smaller run.

CSV comparison and CX reporting

1. Keep the saved input, run ID, collection time, OUTPUT record and COMPANIES record with the full review export. In OUTPUT, inspect failedRequestCount, partial, coverageWarnings, spendingLimitReached and saved counts before calling a sample complete.
2. Export the full default dataset as CSV/JSON, or explicitly select the required fields. The existing Reviews view does not include companyUrl, replyStatus, replyLatencyHours or reviewUpdatedDate; use the full export for identity joins and reply analysis.
3. Group reviews by the company review-page identity, not a mutable company display name. Normalize companyUrl to its /review/<domain-or-slug> path without query parameters when joining to COMPANIES.trustpilotUrl. Retain reviewId for deduplication; do not join on review title or author identity.
4. Compare the sampled ratings and manually reviewed themes with source URLs and sample sizes. Keep overall COMPANIES.overallTrustScore and totalReviewCount separate: a ten-review sample average is not the company's TrustScore or a representative customer-satisfaction estimate.
5. Report only observed reply status. no_reply_observed means no reply text was exposed in this row at collection time, not proof the company ignored the customer. replyLatencyHours compares the exposed review/reply publication dates; it is not a verified first-response SLA. Keep null or reversed/missing timestamp results unknown.

Reporting question	Existing fields	Boundary
Which company and review?	companyUrl, reviewId, reviewUrl, scrapedAt	Retain identity and observation time; deduplicate across exports yourself.
What do sampled reviews rate?	starRating, reviewTitle, reviewBody	Use valid ratings only, disclose sample size and manually assess themes; no automatic sentiment model.
When was feedback posted?	reviewPostedDate, dateOfExperience, reviewUpdatedDate	Posted date and experience date are different; edits can change text or timestamps.
Was a reply visible?	replyStatus, companyReply, companyReplyDate	Missing reply text is only an absence in the observed source.
What timing can be calculated?	replyLatencyHours	Use known, non-negative timestamp pairs; edited/replaced replies are not first-response evidence.
What does the company summary say?	COMPANIES.overallTrustScore, totalReviewCount, rating percentages	Separate current source-level metrics from your limited review sample.
How much was collected?	OUTPUT counts, failures and coverage warnings	A cap, block, filter or partial run must be disclosed; twenty requested rows are not guaranteed.

For repeated snapshots, retain previous exports outside the Actor. Join by reviewId and company identity, then compare known fields while retaining reviewUpdatedDate and both scrapedAt timestamps. A review absent from a later capped sample does not prove deletion. The Actor can save and charge the same review on a later run; deduplicating your report does not refund those events.

A verifiedPurchase value reflects the source's observed verification label, not independent purchase verification. False can also mean no verified label was exposed. Public review text may contain personal information: minimize what you retain or share, and do not publish raw review text as a client example without appropriate rights and review.

See the brand review comparison checklist for an external report handoff. No scored competitor ranking, generated report, consumer profiling or support-system integration is created automatically.

Quick start

Run one company with one recent review:

{
  "companyNames": ["nike.com"],
  "companyUrls": [],
  "maxReviewsPerCompany": 1,
  "sortBy": "most_recent",
  "filterByRating": "all",
  "verifiedOnly": false,
  "proxyConfiguration": { "useApifyProxy": false }
}

Export results as JSON, CSV, Excel, XML, or HTML, or consume them through the Apify API, schedules, webhooks, Make, Zapier, n8n, and other integrations.

What it extracts

Review rows

• Review ID and direct Trustpilot review URL
• Company name and source company page URL
• Star rating
• Review title and review body
• Date of experience and review posted date
• Verified-review flag
• Useful count when shown
• Company reply text and reply date when present
• Observed reply status, reply latency in hours and the review's update date when available
• Scraped timestamp

Company summary rows

Company summaries are written to the current run's COMPANIES key after a valid company page is parsed, even when no reviews match the requested filters. They include:

• Company name, domain, Trustpilot URL, and website URL
• Overall TrustScore and star rating
• Total review count
• Rating distribution percentages
• Claimed status and category when available
• Scraped timestamp

Output dataset

The default dataset is intentionally a clean list of reviews. The Reviews view includes the fields most users need for spreadsheet, BI, API, and sentiment-analysis workflows.

Verified review sample

This shortened sample comes from a successful public Actor run:

{
  "reviewId": "6a2935848b5c27d1ace126be",
  "companyName": "Netflix",
  "companyUrl": "https://www.trustpilot.com/review/netflix.com",
  "starRating": 5,
  "reviewTitle": "Legends",
  "reviewBody": "Legends - just finished the series and WOW! Great true story and amazing acting!",
  "dateOfExperience": "2026-06-10T00:00:00.000Z",
  "reviewPostedDate": "2026-06-10T11:59:32.000Z",
  "verifiedPurchase": false,
  "usefulCount": 0,
  "companyReply": null,
  "companyReplyDate": null,
  "reviewUrl": "https://www.trustpilot.com/reviews/6a2935848b5c27d1ace126be",
  "scrapedAt": "2026-06-10T17:49:53.800Z"
}

Review text, ratings, counts, and reply data can change when Trustpilot updates the page or the reviewer/company edits content.

Input

Field	Type	Default	Description
companyNames	array	["nike.com"]	Company domains or Trustpilot review slugs. The Actor builds the review URL automatically.
companyUrls	array	[]	Full Trustpilot company review URLs.
maxReviewsPerCompany	integer	1	Maximum reviews to save per company. Use 0 only when you intentionally want all available pages.
sortBy	string	most_recent	most_recent, most_relevant, or lowest_rated.
filterByRating	string	all	all, 5, 4, 3, 2, or 1.
verifiedOnly	boolean	false	Save only reviews marked verified by Trustpilot.
proxyConfiguration	object	No proxy	Optional browser proxy settings. The Actor honors your selection and never silently upgrades it to Residential traffic.

Provide at least one company domain, Trustpilot review slug, or full Trustpilot review URL. Duplicate targets are removed before scraping, and each run accepts up to 50 unique companies. Start with one company and review the live price before scaling.

Common workflows

Monitor recent negative feedback

Set sortBy to most_recent and filterByRating to 1 to collect the newest one-star reviews for a brand or competitor.

Compare competitors

Pass several company domains and compare TrustScore, star distribution, review volume, and recent review themes.

Feed sentiment analysis

Export review titles, bodies, ratings, dates, and company replies to a spreadsheet, database, or NLP pipeline.

Build support reports

Use recurring runs to watch new public reviews and route the dataset to a BI dashboard or workflow tool.

Pricing

This Actor uses Pay Per Event pricing.

Event	Price
Actor start	$0.00005 currently; $0.00030 from 24 October 2026 at 02:50 IST. One event through 1 GB, then one per additional GB.
Each successfully saved review-scraped item	$0.001

At the 1 GB default, event charges are:

One company, one run	Current total	From 24 October 2026
1 saved review	$0.00105	$0.00130
25 saved reviews	$0.02505	$0.02530
100 saved reviews	$0.10005	$0.10030
No matching reviews	$0.00005	$0.00030

Apify support confirmed cancellation of the company-check addition previously scheduled for 21 October 2026. It will not take effect. At retained review-only pricing, blocked or empty attempts incur no review or company-check event charge; the small Actor-start charge still applies. No COMPANY_CHECKS ledger is created for the nonexistent fee; backward-compatible billing code uses it only if an active pricing configuration actually registers that event. Standard PPE runs include platform usage; any separate account subscription remains outside these event examples.

Reviews are charged only when they are successfully saved to the dataset. Blocked or empty runs do not charge review-scraped events; the Actor-start charge still applies. The Actor checks its effective event prices and remaining budget before opening another company. All advertised features remain available to free-plan users within their balance and selected spending limit; there is no paid-plan feature gate.

During a platform pricing transition, the code follows each run's active event prices. It does not invent or manually charge an unregistered event. OUTPUT.billing.mode distinguishes the previous review-only pricing from the company-check pricing.

Limits and reliability

• Trustpilot can change page structure or anti-bot behavior.
• Direct browser access is the cost-controlled default. Source blocking can vary by network; select a proxy explicitly only when needed.
• Image, media, and font downloads are blocked and concurrency is bounded. The live cloud default verified on 9 October 2026 is 1 GB; a smaller browser allocation has not been validated.
• Very large runs can take longer because the Actor paginates public review pages.
• Each run accepts at most 50 unique companies and 10,000 reviews per company. Use smaller batches for more predictable completion times.
• Company replies, useful counts, and verification labels are returned only when Trustpilot exposes them on the page.
• lowest_rated uses the one-star filter when no explicit filterByRating is provided.
• A valid company page with no reviews matching the selected filters finishes successfully with an empty default dataset; blocked or unparseable pages still fail visibly.
• Company summaries are stored in the current run's COMPANIES key, while the default dataset stays review-only.

API example

curl -X POST "https://api.apify.com/v2/acts/fascinating_lentil~trustpilot-reviews-scraper/runs?token=YOUR_APIFY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "companyNames": ["nike.com"],
    "companyUrls": [],
    "maxReviewsPerCompany": 1,
    "sortBy": "most_recent",
    "filterByRating": "all",
    "verifiedOnly": false,
    "proxyConfiguration": { "useApifyProxy": false }
  }'

Responsible use

Reply timing and coverage

The live default verified on 9 October 2026 is 1024 MB with one browser page processed at a time. A browser is still required; 256/512 MB is not a validated configuration. Local tests and pricing projections do not establish future customer-run profitability.

The collector skips images, fonts, stylesheets, known analytics hosts and Trustpilot's Next.js UI scripts, while keeping document requests and challenge traffic. It reads the embedded review JSON rather than relying on the interactive page UI. This traffic reduction requires cloud validation; it is not a guarantee of lower cost or successful access on every run.

For additional pages, the candidate implementation derives a JSON pagination URL from the first page's build metadata and retains the requested filters. It uses the existing browser session, validates company identity and review IDs, and falls back to normal navigation after one unsuccessful JSON attempt per company. OUTPUT.jsonPageAttempts and OUTPUT.jsonPagesAccepted expose whether this path worked. Live availability and savings remain unverified; the initial browser page is still required, so this does not solve one-review run costs.

Rows include replyStatus, replyLatencyHours, and reviewUpdatedDate. Reply latency compares the source's review publication and reply publication timestamps; missing or reversed timestamps produce null. It is not a measurement of the company's first response if the source has replaced or edited a reply.

Company summaries are stored in the current run's COMPANIES key, not a shared named dataset. The OUTPUT key reports failed requests and coverage warnings. Pagination stops at repeated pages, 100 pages per company, or a per-company processing safety limit; a partial result is not an exhaustive export.

Use this Actor only for lawful collection of publicly available information. You are responsible for complying with Trustpilot's terms, privacy laws, consumer-review rules, and regulations that apply to your use case.

Do not use the output for spam, harassment, profiling, or unlawful collection of personal data. This Actor is an independent tool and is not affiliated with, endorsed by, or sponsored by Trustpilot.

License

Apache-2.0.