Trustpilot Review Monitor — new reviews & rating-drop alerts

Watch any list of companies on Trustpilot and get only what changed: new reviews (optionally just the 1–2 star ones) and drops in a company's TrustScore. The actor remembers every review it has already reported, so a daily schedule produces a clean alert feed instead of a re-scrape. Reputation-monitoring SaaS charges a monthly subscription for this; here you pay per alert.

Use cases

• Brand reputation: get every new 1–2 star review about your company the morning it appears, with the full text, so support can reply before it spreads.
• Agencies: one monitor per client (monitorId), each with its own webhook into the client's Slack channel.
• Competitor intelligence: watch competitors' review streams and TrustScores; a rating_drop alert is an early signal of a service problem.
• Due diligence / vendor risk: keep a running log of complaints about suppliers, marketplaces or fintechs.

How it works

1. Loads each company's Trustpilot review page (newest first) in a real browser.
2. Compares the reviews against the monitor's memory. Unseen reviews inside your star and language filters are emitted as new_review.
3. Compares the company's current TrustScore with the one stored from the previous run. A fall of at least ratingDropThreshold is emitted as rating_drop with previousTrustScore.
4. Saves the updated memory (up to 50,000 review IDs per monitor) and POSTs a summary to your webhook.

Memory lives in a named key-value store trustpilot-monitor-<hash of monitorId>, so it survives across runs and tasks. Delete that store to reset a monitor.

Input

Field	Default	Notes
companies	["amazon.com","tesla.com"]	Company websites as shown in Trustpilot URLs
minStarsToAlert / maxStarsToAlert	1 / 5	Star band for new_review alerts. 1/2 = negative reviews only
languages	["en"]	Trustpilot language codes; all for every language
maxNewPerCompany	5	Cap on new-review alerts per company per run; the rest come out next run
ratingDropThreshold	0.1	Minimum TrustScore fall (1.0–5.0 scale) that triggers rating_drop
firstRunMode	emitAll	emitAll reports the newest reviews on the first run; baseline records them silently
webhookUrl	""	Optional POST target for the run summary
monitorId	default	One memory per ID — run several watchlists side by side
maxItems	10	Overall cap on alerts per run
proxyConfiguration	Apify residential (US)	Required, see "Why a browser" below

Recommended setup for a daily feed

1. Create a task with your companies, star band and a meaningful monitorId (acme-negative).
2. First run: set firstRunMode to baseline. It records the current reviews and TrustScores and emits nothing, so day one is not a backlog of old reviews.
3. Set maxNewPerCompany and maxItems high enough for a busy day (e.g. 50 / 500) and schedule the task daily. Companies with hundreds of reviews per day (large marketplaces) may deserve every 6 hours; the actor reads at most 100 newest reviews per company and language per run.
4. Point webhookUrl at Slack (incoming webhook), Zapier, Make or your own endpoint.

The default settings ({}) run in emitAll mode so you see real reviews on the first try.

Example: negative-review alerts for two brands

{
  "companies": ["revolut.com", "wise.com"],
  "minStarsToAlert": 1,
  "maxStarsToAlert": 2,
  "languages": ["en", "de"],
  "maxNewPerCompany": 50,
  "maxItems": 500,
  "ratingDropThreshold": 0.1,
  "firstRunMode": "baseline",
  "webhookUrl": "https://hooks.slack.com/services/XXX/YYY/ZZZ",
  "monitorId": "fintech-negative"
}

Output

One record per alert. changeType is new_review or rating_drop.

{
  "company": "amazon.com",
  "companyName": "Amazon",
  "trustScore": 1.6,
  "reviewCount": 49213,
  "reviewId": "68d9573bf1ddb15da6e70cbc",
  "rating": 1,
  "title": "Package never arrived",
  "text": "Ordered on the 20th, tracking said delivered but …",
  "reviewerName": "Jane D",
  "reviewerCountry": "US",
  "reviewDate": "2026-09-28T17:10:10.000Z",
  "experienceDate": "2026-09-27T00:00:00.000Z",
  "isVerified": false,
  "replyText": null,
  "replyDate": null,
  "reviewUrl": "https://www.trustpilot.com/reviews/68d9573bf1ddb15da6e70cbc",
  "changeType": "new_review",
  "previousTrustScore": null,
  "firstSeenAt": "2026-09-28T19:02:11.412Z",
  "monitorId": "default"
}

A rating_drop record carries the company fields, trustScore (current), previousTrustScore and reviewUrl pointing at the company profile; the review fields are null.

Webhook payload

POSTed once per run as application/json, also saved as SUMMARY in the run's key-value store:

{
  "monitorId": "fintech-negative",
  "runAt": "2026-09-29T06:00:03.118Z",
  "newReviewCount": 7,
  "ratingDropCount": 1,
  "recorded": 0,
  "seenTotal": 412,
  "baseline": false,
  "companies": { "revolut.com": { "trustScore": 4.2, "reviewCount": 210331, "at": "2026-09-29T06:00:03.118Z" } },
  "records": [ { "...first 50 records, same shape as the dataset..." } ]
}

Why a browser (and 4 GB of memory)

Trustpilot gates its pages behind a "Verifying Connection" check that blocks plain HTTP clients and headless browsers. This actor runs a real, headed Chromium on an Apify residential proxy, which passes the check. Run it with 4096 MB of memory; a browser session is why the start fee is a little higher than for a plain API monitor.

Pricing

Pay per event: a small start fee plus a fee per alert emitted. A daily run that finds nothing new costs only the start fee.

Notes

• Reviews outside the star band are remembered too, so widening the band later does not replay old reviews.
• Companies without a Trustpilot profile log a warning and are skipped.
• The first run without state is the only one where firstRunMode matters.