Frequently asked questions

The Trust Score evaluates credibility signals in the reviews currently available for analysis. It does not verify product quality, safety, authenticity, seller legitimacy, durability, or suitability for you. Read it alongside the review count, the dimension breakdown, any product recalls, the return policy, and independent evidence.

A star rating records how shoppers scored the product. The Trust Score assesses how credible the set of reviews looks as a body of evidence. A product can hold a high star rating built on reviews that show manipulation signals, and a product with an average star rating can have reviews that look entirely genuine. The two answer different questions, so they often disagree.

The fewer reviews there are, the less confidence any conclusion carries. Analyses work best when a product has at least 20 reviews. Below that, a handful of unusual reviews can dominate the outcome, so treat the score as a weak signal rather than a verdict, and always read it together with the review count.

We detect patterns associated with both, but we cannot prove that any single review was written by AI or paid for. A review written in exchange for a refund or a free product comes from a real buyer who really received the item, so the text itself may look completely ordinary. Treat our signals as evidence to weigh, not as proof about any individual review.

Retail listings often pool reviews across colours, sizes, models, and older and newer versions of the same product. When that happens, some reviews in the analysis describe an item that differs from the one you are looking at. This can move the result in either direction, and we cannot tell which variant each review was written about.

Every report shows the date it was generated. It reflects the reviews that were visible on that date and does not update on its own as new reviews appear. If a product has seen significant activity since then, run the analysis again for a current view.

In short: we score a product’s review set on several dimensions — how well the written text matches the star ratings, how specific and evidence-based the reviews are, how the ratings are distributed, and how much promotional language appears — then combine them into one Trust Score. Trust in Reviews walks through the method and what each dimension can and cannot show. Most analyses finish in about 30 seconds.

Fakespot closed on 1 July 2025, after Mozilla acquired it. ReviewMeta stopped operating before that. ReviewDetector is still running and still free, which is why people searching for a Fakespot alternative in 2026 end up here. Paste a product link to get a Trust Score and a dimension breakdown for Amazon, Best Buy, Target and TikTok Shop listings.

The usual reasons are an unsupported platform, a listing restricted in your region, too few reviews to work with, a page we cannot reach, or a link that is not a direct product URL. If your link points to a TikTok Shop seller rather than a product, use the TikTok Shop Legit Checker instead.

An analysis runs on the publicly visible reviews of the listing you submit. For what we collect, how long we keep it, and how to reach us about your data, see our Privacy Policy.

Price alerts let you follow a product and hear from us when its price falls. They are a secondary feature — review analysis is what ReviewDetector is built for.

Alerts are intended to tell you about a meaningful price drop rather than to notify you on every fluctuation.