Review Sentiment Scorer for Trustpilot Exports

Review Sentiment Scorer turns customer reviews exported from Trustpilot (or any review platform) into a sentiment-ready dataset. Paste reviews as JSON or CSV, and get for each one a sentiment score from -1 to 1, a positive / neutral / negative label, the topics mentioned (delivery, price, support, quality, refund, website) and the words that drove the score. It does not visit Trustpilot or any website to collect reviews: you bring the text, it does the analysis, so it runs in seconds and never gets blocked.

What can this Actor do?

• Score every review with a transparent, rule-based English sentiment model (negation and intensifiers handled), blended with the star rating when you provide one.
• Flag rating mismatches: 5-star reviews with angry text, or 1-star reviews with praise. These are the reviews worth reading first.
• Tag aspects (delivery, price, support, quality, refund, website_app) so you can see what customers complain about.
• Export results as JSON, CSV, Excel or via the Apify API, Make, Zapier and n8n, ready for dashboards and spreadsheets.
• On the Apify free plan you can score thousands of reviews with your monthly credits.

How to use it

1. Click Try for free and open the Input tab.
2. Put your reviews in Reviews (JSON list), paste a CSV in CSV text, or give a public CSV URL. A CSV needs a header row with columns such as title, text (or content), rating, author, date.
3. Set the maximum number of results and click Start.
4. Download the dataset from the Output tab.

Input

{
  "reviews": [
    {"title": "Great service", "text": "Fast delivery and very helpful support. I love it!", "rating": 5, "author": "Anna"},
    {"title": "Never again", "text": "Terrible experience, parcel arrived damaged.", "rating": 1}
  ],
  "neutralThreshold": 0.1,
  "maxItems": 100
}

Output

{
  "id": "1",
  "author": "Anna",
  "rating": 5,
  "title": "Great service",
  "text": "Fast delivery and very helpful support. I love it!",
  "sentimentScore": 0.9,
  "sentimentLabel": "positive",
  "ratingMismatch": false,
  "aspects": ["delivery", "support"],
  "sentimentKeywords": ["love", "helpful"],
  "wordCount": 11
}

How much does it cost?

This Actor uses pay per event pricing: you pay a small amount per scored review (one result event), and nothing for rows that are empty. 1,000 scored reviews cost the price of 1,000 result events shown in the Pricing tab. There is no proxy and almost no compute, so runs stay cheap.

FAQ

Does it scrape Trustpilot?

No. It analyses review text that you provide, for example from your own Trustpilot business export or any other source you are allowed to use. This avoids anti-bot walls entirely.

Which languages are supported?

English only for now. Other languages get a score based on stars only when a rating is given, otherwise neutral.

How accurate is it?

It is a fast lexicon-based model, good for triage and trends, not a replacement for human reading. Sarcasm can fool it; the rating mismatch flag helps spot such cases.

Is it legal to use?

You are responsible for the data you provide. Reviews can contain personal data (names): handle them in line with the GDPR and remove what you do not need.

Can I use it through the API?

Yes. Every run can be started and read through the Apify API and its Python and JavaScript clients.

---

Created with the help of artificial intelligence.