How to scrape Trustpilot reviews without a business account, an app token or a login
Last updated
Maintained by Datapika · We publish this product under the openclawai account on Apify. Product reference
The Datapika Trustpilot scraper runs on Apify and takes a website domain, a company name or a Trustpilot review page URL. It returns every review as one row, with star rating, title, text, language, dates, verified flag, reviewer country and the company reply, plus a free profile row with TrustScore, rating distribution and reply rate. Trustpilot's pages stop at 200 reviews per filter; the deep sweep combines star, date and language streams and recovered 1,500 Airbnb reviews in 320 seconds on September 7, 2026. Price is $0.005 per review, with no login, cookie or token.
Why is Trustpilot hard to scrape, and how does this actor read it?
Trustpilot sits behind a managed bot-protection layer: a plain HTTP client, a headless browser with default settings, and even a residential IP with a real browser fingerprint receive a JavaScript challenge instead of the page. The review data itself is public, embedded in each page as structured JSON for the browser to render, and needs no account to view. What blocks most scrapers is the challenge, not the data.
The actor reads the pages through Apify's anti-bot proxy, which solves the challenge server-side, then parses the embedded JSON rather than the HTML, so field names stay stable when Trustpilot changes its layout. The proxy is included in the per-review price. No Trustpilot Business account, API key, mobile app token or browser cookie is used, which matters because several other scrapers depend on a captured app token that expires and stops every run until someone replaces it.
- Data source: the structured JSON Trustpilot embeds in every public review page, not scraped HTML
- Bot wall: solved by an anti-bot proxy included in the price, no cookie reuse, no browser to maintain
- Inputs: a domain such as www.notion.so, a company name such as Airbnb, or a review page URL
- Nothing to paste: no Business account, API key, app token or login
- Company-name inputs are matched against Trustpilot search; an exact name wins, then the most-reviewed profile
How do you get more than 200 reviews per company?
Trustpilot serves at most 10 pages of 20 reviews for any one combination of filters; page 11 comes back empty even when the profile lists 100,000 reviews. The deep sweep, on by default, works around that by requesting many filter combinations, each a fresh 200-review stream: every star rating, every date bucket from the last 30 days to all time, and every language the company has more than 20 reviews in, then removes duplicates by review ID. Streams are fetched by a work queue that keeps 16 threads busy, and reviews are written to the dataset every 200, so a long run shows progress and a timeout keeps what was fetched.
On September 7, 2026 a 1,500-review pull of www.airbnb.com finished in 320 seconds and a 10-company run at 100 reviews each in 460 seconds. Brands with tens of thousands of reviews return several thousand rather than all of them, because the number of distinct streams is finite; the free company row carries Trustpilot's own numberOfReviews so you can see the coverage you got.
- Hard limit: 10 pages of 20 reviews per filter combination, on the website and for every scraper
- Deep sweep: star x date bucket x language streams, deduplicated by review ID
- Measured September 7, 2026: 1,500 Airbnb reviews in 320 s; 10 companies x 100 reviews in 460 s
- Reviews flush every 200 rows, so partial results survive a timeout
- numberOfReviews on the company row shows Trustpilot's own total for coverage checks
What fields does each review and company row include?
Review rows carry dataType review, id, url, company, companyDomain, companyUrl, rating from 1 to 5, title, text, language, publishedAt, experiencedAt, updatedAt, isVerified and verificationLevel, source (organic or invited), likes, reviewerName, reviewerId, reviewerCountry, reviewerReviewCount, reviewerReviewsOnThisCompany, hasReply, replyText, replyPublishedAt, isPending, location, and the sourceInput that produced them.
Company rows, one per company and free, carry name, domain, website, trustScore, stars, numberOfReviews, numberOfReviewsLast12Months, ratingDistribution as counts per star, reviewsByLanguage, categories, primaryCategory, isClaimed, isClosed, countryCode, email, phone, address, replyPercentage and averageDaysToReply. Server-side filters for stars, dateRange, languages, verifiedOnly and withReplyOnly are sent to Trustpilot, so a 1-star-only run fetches, and bills, only 1-star reviews.
- Review: rating, title, text, language, publishedAt, experiencedAt, isVerified, source, likes
- Reviewer: reviewerName, reviewerCountry, reviewerReviewCount, reviewerReviewsOnThisCompany; public display fields only
- Reply: hasReply, replyText, replyPublishedAt
- Company, free: trustScore, numberOfReviews, ratingDistribution, reviewsByLanguage, replyPercentage, averageDaysToReply, contact fields
- Filters applied at the source: stars, dateRange, languages, verifiedOnly, withReplyOnly
How much does a Trustpilot scraper cost on Apify?
The table records store figures as shown on September 7, 2026. memo23, the largest with 2,472 users and a 4.80 rating from 20 reviews, lists from $0.50 per 1,000 results plus paid AI analysis events. automation-lab/trustpilot, 1,963 users, lists a $0.005 start fee per run plus $0.575 per 1,000 reviews. theagents lists $0.50 per 1,000 reviews with a 5.0 rating from 7 reviews. zen-studio lists $1.99 per 1,000 but showed a 39.7 percent failure rate over the last 30 days on its store page.
Datapika charges $5 per 1,000 reviews, a single event with the anti-bot proxy included, the company profile row free, and no start fee. It is the most expensive per row in the table; what you pay for is a run that needs nothing pasted in, filters applied at the source so you only pay for reviews you asked for, and a deep sweep that goes past the 200-review wall.
- memo23/trustpilot-scraper-ppe: from $0.50 per 1,000 results, 2,472 users, 4.80 rating from 20 reviews
- automation-lab/trustpilot: $0.005 start + $0.575 per 1,000 reviews, 1,963 users
- theagents/trustpilot-reviews: $0.50 per 1,000 reviews, 557 users, 5.0 rating from 7 reviews
- zen-studio/trustpilot-review-scraper: $1.99 per 1,000, 39.7 percent of runs failed in the last 30 days
- Datapika: $5 per 1,000 reviews, proxy included, company row free, no start fee, published September 2026
| Actor | Listed price | Users | Rating (reviews) | Notes |
|---|---|---|---|---|
| Datapika Trustpilot Reviews Scraper | $5 / 1,000 reviews | 1 | new (Sept 2026) | No token, anti-bot proxy included, deep sweep past 200, free company profile row |
| memo23/trustpilot-scraper-ppe | from $0.50 / 1,000 results | 2,472 | 4.80 (20) | Paid AI analysis events at $0.02 and $0.05 |
| automation-lab/trustpilot | $0.005 start + $0.575 / 1,000 reviews | 1,963 | 5.0 (1) | Start fee per run |
| theagents/trustpilot-reviews | $0.50 / 1,000 reviews | 557 | 5.0 (7) | |
| zen-studio/trustpilot-review-scraper | $1.99 / 1,000 reviews | 659 | none | 39.7% of runs failed, last 30 days |
Sources you can inspect
Product documentation and implementation references. These references describe supported options and billing, not independent proof of performance. Confirm current details before running; examples and estimates are not guarantees.
How to do it
- 1.Open https://apify.com/openclawai/trustpilot-reviews-scraper and add companies one per line: a domain such as www.amazon.com, a name such as Notion, or a review page URL.
- 2.Set maxReviews per company (up to 20,000) and any filters: stars such as 1 and 2, dateRange such as last30days, languages such as de, verifiedOnly or withReplyOnly.
- 3.Leave deepSweep and includeCompanyInfo on. Run it; reviews land every 200 rows and each company adds one free profile row.
- 4.Export JSON or CSV, or schedule a daily run with dateRange last30days and diff review ids to catch every new review the day it lands. Call it by API or MCP for agents.
Questions, answered
Do I need a Trustpilot account or API key to scrape reviews?
No. The actor reads Trustpilot's public review pages, which need no login, through an anti-bot proxy included in the price. Nothing is pasted in and nothing expires. Trustpilot's own APIs are issued to Trustpilot Business customers through their account, and cover a business's own reviews rather than any company's.
How much does it cost to scrape 1,000 Trustpilot reviews?
$5, as one event at $0.005 per review row. Company profile rows, hint rows and error rows are free, there is no start fee, and the proxy is included. A 200-review pull for one company costs $1 and takes about a minute; a 1,500-review deep sweep cost $7.50 and took 320 seconds on September 7, 2026.
Why does the run return fewer reviews than the company's total?
Either your cap or filters limited it, or the number of distinct filter streams ran out: Trustpilot serves at most 200 reviews per combination, and very large brands have more reviews than the star, date and language combinations can reach. The company row's numberOfReviews is Trustpilot's total for reference.
Can I get only 1-star reviews, or only reviews with a company reply?
Yes. stars, dateRange, languages, verifiedOnly and withReplyOnly are sent to Trustpilot as filters, so the run fetches only matching reviews and bills only those. A multi-star filter such as 1 and 2 is supported in one run.
Why does verifiedOnly return so few reviews?
Because Trustpilot marks few reviews as verified. On September 6, 2026, Booking.com showed 34 verified reviews out of about 126,000, and exactly 1 that was both verified and replied to. The actor returns what the site has; the filter is not broken.
Which reviewer details are included?
The public display fields Trustpilot shows on the page: display name, country, number of reviews written and the verified flag. No email addresses or private data are collected.