Two public replies ready to paste, written for the buyer who will read them later and reviewed against Amazon's communication policy before you see them.
The problem with responding to a review isn't finding the words: it's that a single extra phrase — "if you're satisfied, please update your rating" — could cost you your account.
You paste the review as is, set the stars, and before you use a single AI call, the app has already done deterministic work: it detects the review's language, the reason (use, defect, missing piece, size, price, expectation, shipping…), the severity, and what approach is appropriate. If there's a potential security risk, an accusation of counterfeiting, or insults, it informs you and reminds you that this should be reported to Amazon, not discussed publicly.
Then it drafts TWO variants, a brief one of two or three sentences and a complete one of four or five, both with the same fixed structure: gratitude and acknowledgment of the problem without admitting any legal fault that hasn't been proven, genuinely useful information (usage instructions, specifications, clarifications), an offer of help ONLY through Amazon's official channels, and a closing with your signature. It responds in the language of the review and does not reproduce any personal data detected in the text: order number, email, or phone number are excluded.
And here’s the part that a generic chat doesn’t do: no variant is delivered without passing through the communication policy validator and the four risk lexicons. What the server marks as a violation is not delivered to you as a valid response: it is regenerated, up to two more times, with the explicit instruction to remove what failed. If it continues to fail on the third attempt, it delivers it MARKED, with the flagged phrase and a notice that the decision is yours, never passing off as acceptable something that is not.
And if you rewrite it manually — which is when the real risk appears, because the prohibited phrase is almost always added by the seller — when saving the edit, it goes through the same validator on the server again. The checker on your own text, moreover, is free and unlimited and does not require an account.
Use it free without an account. If you want to save your result, you can sign up afterwards.
Save / Sign upWritten for those who haven't purchased yet
the public reply is not drafted to calm the complainer, but so the next buyer sees service and resolves their doubt. That is the criterion for all the text.
Two variants, brief and complete
one of two or three sentences for reviews that only need clarification and another of four or five when there’s a need to explain a usage or specification. You choose, edit, and copy.
Fixed four-part structure
gratitude and acknowledgment without admitting legal fault, useful information, offering help through the official Amazon channel, and closing with your signature. No links, no emails, no phone numbers, and no mention of any channel outside of Amazon.
Nothing that violates policy reaches your clipboard as compliant
each variant goes through the communication validator and the four risk lexicons. What is marked as blocked regenerates up to two times; if it still fails, it is delivered marked with the indicated phrase, never disguised as a safe response.
The five red lines that are monitored
asking, altering, or implying a change in the review; conditioning a refund, replacement, or gift to what happens with it; diverting the conversation outside of Amazon; disqualifying the buyer; and sneaking in prohibited promotion. Plus the detection of personal data in the text.
Ventaja 6
Responds in the language of the review, automatically detected among Spanish, English, German, French, Italian, and Portuguese. If you sell in multiple marketplaces, you don’t have to translate anything manually.
Before AI, the free diagnosis
reason for the review, severity, and recommended approach are calculated instantly on the server at no cost. And if the review is an insult or an accusation of falsification, it tells you and recommends reporting it instead of responding.
Your editing is also reviewed
when you rewrite the response manually, upon saving it goes through the validator again. And the policy checker on your own text is free, unlimited, and without registration, even if you don’t generate anything.
The first response is free and no registration is required, and the policy checker on your own text never runs out.
In the features that use AI, responses are generated by artificial intelligence and are not a substitute for professional advice.