If you used to lean on Fakespot or ReviewMeta to sanity-check an Amazon listing before buying, you've probably noticed both are effectively gone. Mozilla switched off Fakespot's Firefox Review Checker on June 10, 2025, and its extensions, apps, and website stopped working on July 1, 2025, after Mozilla decided the standalone product no longer fit a model it could sustain. ReviewMeta, the other free name most shoppers recognized, is also no longer a working option: its founder publicly announced he was stepping away and looking for a successor, and the site and extension have since gone quiet and stopped pulling fresh data. That leaves a real gap — a lot of people lost their one-click trust check around the same time.
The good news is that you don't actually need to read a wall of reviews — or a third-party tool's summary of the review text — to catch most manipulation. The strongest tells are structural, and they sit right on the product page in plain sight: the shape of the star distribution, how many ratings there are, what share are verified purchases, and when the reviews were posted. This guide teaches you to read those signals yourself, and then shows where Ryohin Checker automates the same check for free when you'd rather paste a link and get a verdict.
For most of the past decade, the standard advice for a skeptical Amazon shopper was "run it through Fakespot" or "check it on ReviewMeta." Both tools graded a listing and gave you a quick read on whether the reviews could be trusted. In 2025 that workflow broke.
Mozilla, which had acquired Fakespot in 2023, shut down the built-in Firefox Review Checker on June 10, 2025, and made the Fakespot extensions, mobile apps, and website unusable as of July 1, 2025, saying the standalone product no longer fit a model it could sustain. ReviewMeta's path was different but the outcome is similar: its founder posted publicly that he was stepping back and seeking a successor to keep the project alive, and in the time since, the site and extension have gone dark and stopped updating. There's no confirmed permanent-shutdown date, but for practical purposes it's no longer a usable tool.
The result is a gap. The two most recognized free review checkers are both effectively gone, and the demand didn't disappear with them. If anything, fake reviews stayed in the headlines: the U.S. Federal Trade Commission's Rule on the Use of Consumer Reviews and Testimonials — which bans buying, selling, and AI-generating fake reviews — took effect on October 21, 2024, banning the buying, selling, and AI-generation of fake reviews. The problem is alive and regulated; the easy consumer tool just left the building.
The instinct when you suspect fakes is to start reading reviews, hunting for stilted phrasing or suspiciously glowing language. That's the weakest possible approach in 2026, for two reasons.
First, text is the easiest thing to fake. AI can now generate fluent, specific, totally natural-sounding five-star reviews at scale, so "it reads fine" tells you almost nothing. The old red flags — broken English, copy-paste phrasing — still exist, but the convincing fakes sail right past them.
Second, the review bodies are increasingly hard to even get at. Amazon has tightened access to review text in recent years: logged-out review pages often return errors, and full review text is harder to pull from the page than it used to be. That's exactly why text-scraping tools have struggled and why it's risky to build anything on harvesting full review bodies.
Structure is different. To move a product's rating, a seller has to manufacture volume, and volume leaves fingerprints in the aggregate numbers that are far harder to disguise. A burst of paid reviews distorts the star distribution. It tends to arrive in a tight window of dates. It often skips the verified-purchase step or relies on reimbursed buys. It can push the rating count and the average out of natural proportion. You can see all of that without reading a single sentence — and so can an automated checker.
Here are the signals worth learning. Each one is visible on the listing itself or in its rating breakdown, and each is hard for a manipulator to hide because it's a side effect of how fake volume gets created.
None of these is a smoking gun on its own. A single flag can have an innocent explanation. The point is to weigh them together: one yellow flag is noise, three stacked together is a pattern.
You can run this in under a minute on any listing. No tools required — just the product page and its rating breakdown.
Work through it in order. By the end you'll have a feel for whether the numbers hang together or whether something was pushed.
Doing the self-check by hand is great for one or two products. When you want speed — or a second opinion that's consistent every time — that's the job Fakespot and ReviewMeta used to do, and the gap Ryohin Checker is built to fill.
Ryohin Checker is a free, on-demand tool. You paste an Amazon product link and it reads the same public, structural signals you'd check by hand — the star distribution, the number of ratings, the verified-purchase share, and the clustering of posting dates — and returns a trust score out of 100 with the specific reasons behind it. It deliberately does not depend on scraping full review text, which is exactly why it keeps working where text-based tools broke. There's a JP edition for amazon.co.jp and an EN edition for amazon.com.
One thing that sets it apart: it only puts its name behind products that pass. A product makes the recommended list only when its trust score clears 75, its average is at least 4.0 stars, and there's enough data to judge confidently. Everything else is simply left off the list — the tool never trashes a product or accuses a seller; it just declines to vouch. If you'd rather skip the checking entirely and start from a shortlist that already cleared the bar, browse the picks that passed.
Paste a link to check a product, or browse the curated rankings:
No tool — and no human reading a page — can declare a review fake with certainty from the outside. What structural analysis gives you is a well-reasoned probability, not a verdict. A clean-looking listing can still hide manipulation, and an unusual-looking one can have a perfectly innocent story (a viral moment, a big launch, a genuinely polarizing product).
Ryohin Checker is built around that honesty. The score is a risk estimate from public structure, the reasons are shown so you can judge them yourself, and a low score is framed as "we can't vouch for this," not "this is fraud." Where the data is thin, it says so rather than guessing.
Use it the way you'd use any single signal: as one input into your own judgment. Cross-check a big purchase against the seller's track record, the product's history, and your own read of the page. The goal isn't to outsource the decision — it's to make a faster, better-informed one now that the old one-click tools are gone.
In May 2026, the AI-detection firm Pangram Labs ran 30,000 front-page reviews from 500 of Amazon's best-selling products through its detector. About 3% — 909 reviews — were flagged as AI-generated with high confidence. That may sound small, but these were the reviews shoppers see first, on the products that sell the most.
Two findings are worth keeping in mind the next time you read a listing. First, the AI-written reviews leaned far more positive than the human ones: 74% gave a 5-star rating, versus 59% of genuine human reviews. Second — and this is the part that surprised people — 93% of those AI reviews carried the "Verified Purchase" badge.
This is exactly why we lean on structure rather than wording. A fluent, glowing, "verified" review can still be synthetic, and you can't tell by reading it. What the badge and the prose can't easily hide is the shape of the data underneath: how the stars are distributed, how fast they arrived, and whether the count makes sense for the product's age.
You're not the only line of defense. In its first Trustworthy Shopping Experience Report (2025), Amazon said it blocked hundreds of millions of suspected fake reviews and, through legal action, helped shut down more than 100 websites that brokered fake reviews. Much of this now runs on AI and machine-learning checks that screen reviews before they're published.
That's genuinely good news, and it means most obvious manipulation never reaches you. But enforcement and evasion move together: the same study above shows AI-written reviews still slip onto front pages. Amazon's filters and a quick structural self-check aren't competing — they're layers. Ryohin Checker is meant to be the second layer, helping you sanity-check what survived the first.
As an Amazon Associate, Ryohin Checker earns from qualifying purchases. Verdicts are estimates inferred from public page data (star distribution, number of ratings, posting dates, verified-purchase share) and do not guarantee authenticity (mistakes are possible). We do not store or republish review text. Rankings and recommendations are not influenced by commissions.