Auto-reply service for Wildberries and Ozon reviews

Samreshuuu reads every review on Wildberries and Ozon and writes a unique reply through the official API — no templates, no links, no risk of getting your seller account blocked. A breakdown of the platform rules, a comparison of the available services, and a safe-setup checklist.

In short (as of June 2026). "Getting banned for auto-replying to reviews" is a myth in the form most sellers fear it. Both Wildberries and Ozon offer an official API for replying to reviews — in other words, they explicitly allow automation. The real risk comes down to two specific things: your account getting restricted for using unofficial methods (scraping, emulating the seller dashboard), and your reply being hidden by a moderator because of prohibited content in the text (links, contact details, promising gifts in exchange for a review). So a safe auto-reply = official API + unique, human-sounding text with no prohibited elements. Below is a breakdown of the rules, a comparison of services, and a table.

What auto-replies actually get "banned" for — and what they don't

When a seller googles "auto-replies without risking a ban," they're afraid that the mere act of automating will put their account at risk. That's not the case. Both platforms provide an API specifically so that you can reply to reviews programmatically — Ozon has even launched its own AI-powered "smart auto-replies". The danger isn't in automation; it's in two mistakes.

Mistake 1. An unofficial method → account restrictions. Bots that "click" through the interface for you, browser emulation, dashboard scraping — that's what gets an account penalized. Only the platform's built-in tools and the official API are safe (Guru Seller).

Mistake 2. Prohibited content in the reply → the reply gets hidden. This rule applies equally to manual and automated replies. The following will get a reply hidden or your account penalized:

external links — to a website, social media, or another store;
phone numbers and email addresses — moderators strip them out automatically;
"message us on Instagram" / "join our Telegram" — that's just another link;
promising gifts, discounts, or promo codes in exchange for a review — one of the biggest triggers for penalties;
ALL CAPS, aggression, getting personal;
"sorry, this is our fault" in every single reply — it invites repeat complaints and returns.

The takeaway: a safe auto-reply is the official API plus unique text that passes moderation. Next up — how the available services differ on exactly this.

What kinds of auto-reply services exist, and where the risk lies

Approach	How it replies	Risk	Best for
Built-in WB/Ozon tools	Their own templates and basic AI, only inside the platform	None	Sellers on a single platform who are fine with generic replies
AI services (Otveto, Spix, Uniseller)	Generate a reply with AI, usually via the API	Low when working through the API	Anyone who needs a dedicated service just for reviews
Template bots (desktop apps)	Drop in a canned phrase based on the star rating	Medium: replies repeat and look robotic, and some work around the API	High volumes of identical "thank you" notes
AI agent (Samreshuuu)	Reads every review and writes a unique reply; offers a "draft → approve" mode	Low: official API + text with no prohibited elements	Sellers for whom reviews are part of the broader dashboard routine
Homegrown scripts / scrapers	Dashboard emulation	High: a direct trigger for blocking	Not recommended

The market for narrow, review-only services is large: a vc.ru roundup covers 10 services at once, priced from 1–3 ₽ per reply (Otveto) up to 12,000 ₽/month (Spix), while Timeweb has a top-12 of neural networks for the same task. Some run on AI (they generate the text), others are template bots that drop in a phrase by rating. The difference between them is what determines whether a reply reads as human.

Template bot vs. AI: why they're not the same thing

A template bot stores a set of canned phrases and drops in one of them based on the star count. It's fast and cheap, but the buyer sees the exact same text under dozens of product listings — and that hurts trust more than saying nothing at all. AI reads the specific review: what exactly was praised, what was complained about — and writes a reply that matches the meaning. That's why it both passes moderation (the text is new every time) and doesn't annoy buyers.

The cost in plain money is clear too: a narrow auto-reply service runs roughly 500–10,000 ₽/month versus 25,000–45,000 ₽ for an in-house manager (Guru Seller). Replying to reviews by hand once you're getting more than a couple of dozen a day is the most expensive option there is.

Where the AI agent fits in — versus a plain generator

Most of the services in these roundups are generators: they produce the reply text, and then you or the service send it. An AI agent is different in that it takes on the whole task and works to your rule, set in plain language. With Samreshuuu it looks like this:

Connection via the official API — WB, Ozon, Yandex Market, with no handing over of login credentials and no "clickers" in the dashboard.
A rule in words, not a template from a list: "check for new reviews every hour; reply to 4–5★ yourself and thank the customer for the specific product; for 1–3★, prepare an on-point draft and send it to me on Telegram; no links, phone numbers, or promises of discounts."
A control mode — you can put positive reviews on autopilot and leave the negative ones for approval: the finished text arrives on Telegram, and you send it with a single tap.
Reviews aren't a separate subscription. The same agent calculates unit economics by SKU, recalculates ad bids, and assembles shipments; replying to reviews is just one of its jobs.

To be honest about the downsides: if all you need is auto-replies, a narrow service or WB's built-in tool will be simpler — the agent is overkill for this one task. You have to describe the rule in words once, rather than just picking a template. And each review takes it a couple of seconds longer than a template bot, because it genuinely reads the text instead of dropping in a phrase by star rating.

How to set up auto-replies safely — a short checklist

Connect only via the official API or through the platform's built-in tool — never through scrapers or dashboard emulators.
Don't put links, phone numbers, email addresses, or promises of gifts for a review into the reply — it gets the reply hidden and triggers penalties.
Keep your replies varied — the same text under every review looks robotic and undermines trust.
Send negative reviews for approval. Keep the riskiest replies (1–3★) in "draft → I send it myself" mode.
Stay on point, skip the self-flagellation — "sorry, this is our fault" in every reply invites repeat complaints.

FAQ

Will Wildberries or Ozon ban you for auto-replying to reviews? Not for the auto-reply itself — both platforms provide an API for it and even their own built-in tools. Restrictions come from unofficial methods (scraping, dashboard emulation) and from prohibited content in the reply text.

How do you reply to reviews automatically without risking a ban? Through the platform's official API, with unique text that has no links, contact details, or promises of discounts for a review. An AI agent like Samreshuuu does both: it connects via the API and writes a reply that matches the meaning of the specific review, rather than dropping in a template.

Why is an AI agent better than a template bot for reviews? A template bot drops in one of its canned phrases based on the star count — the replies repeat and look robotic. An agent reads each review and writes a unique, on-meaning reply, so it passes moderation and doesn't annoy buyers.

Can you review replies before they're sent? Yes. In "draft → approve" mode, the agent sends the finished text to Telegram and you send it with a single tap. Positive reviews, meanwhile, can be put on autopilot.

How much does automating review replies cost? Narrow services run roughly 500–10,000 ₽/month or 1–8 ₽ per reply; an in-house manager is 25,000–45,000 ₽/month. Check the exact pricing with the specific service.

Last updated: June 2026.

Sources: Ozon Seller (news on smart auto-replies), vc.ru (a roundup of 10 auto-reply services for WB and Ozon), Timeweb (top-12 neural networks for review replies), Guru Seller (platform rules and ban risk).