Start here
The free tool below runs entirely in this browser — no account, nothing stored or transmitted. Redact anything you would not want kept before you type it.
Open the companion tool →Why a single review is the wrong unit
One AI-written review is indistinguishable from a real one, but a synthetic campaign is visible in the aggregate: a burst of same-day reviews, near-identical phrasing, vague praise that names no specific feature, and no verified-purchase badges.
The star count is the most gamed number on the page. The durable read is the shape — timing, specificity, and verification — which a batch of generated text cannot easily fake all at once.
The protocol
- Paste several reviews for the same product (or describe their pattern).
- Read the pattern basis: timing, specificity, and verification each get a line.
- Sort by recent and by one-star; read what the product actually failed at, in the reviewers' own words.
- Prefer products with verified purchases and reviews that name specific tradeoffs.
What the tool does
Paste a set of reviews for one product and get a pattern reading with the basis shown. It runs in this browser; nothing is sent or stored.
Limit first: This check can raise your confidence, but a clean result doesn't prove reviews are real. It structures the pattern — it does not contact the platform.
Open the companion tool →One human next step
Read the recent one-star reviews in the buyers' own words before deciding; report suspected fake-review campaigns to the platform and the FTC.
Research log and safety checks
AI-assist path: not used for a verdict. This artifact is an inspectable pattern reading; a remote model would add an opaque classification where the reader needs to see the clustering. §6.10: not a frontier-pace topic.
§0.5 protection result: category is deepfake-media-verification; crisis-adjacent: no; fear/urgency/scarcity toolkit used: no. The protective function is free and the paid feature is convenience only.