AI bots already read your store.
See exactly who, and take control.

RobotsLLM shows every AI bot and every ChatGPT referral down to the order, manages access with diff and rollback, and makes your catalog readable for AI. No extra server.

llms.txt
# akumulyator.center
# Generated by RobotsLLM 0.1.0

> Інтернет-магазин акумуляторів, ДБЖ та
> сонячних електростанцій в Україні.

## Категорії
- [Автомобільні акумулятори](/avtomobilni-akumulyatori.md)
- [Гібридні інвертори](/invertori/gibridni-sonyachni-invertori.md)
- [Сонячні панелі](/sonyachni-stancii/sonyachni-paneli.md)
- [ДБЖ та резервне живлення](/rezervne-zhivlennya-bezperebiyniki.md)

## Політики
- [Доставка і оплата](/delivery.md)

Measured on a live store, not in a deck

We installed RobotsLLM on our own store akumulyator.center. Here are 27 days of its AI analytics, collected by the add-on itself (16 August - 12 September 2026).

279 289

bot requests recorded

IP and rDNS bot verification

17 808

times ChatGPT opened store pages on user requests

58

visits from people coming from AI assistants

ChatGPT 76 · Gemini 19 · Perplexity 2

3

orders attributed to AI sources

Stream of AI bots hitting an online store server

45% of all bot requests came from training bots, 114,199 of them from a single Meta bot. Without control you simply pay for that with your server.

Shoppers ask ChatGPT. What does it know about your store?

AI answers without you

When a shopper asks an assistant "which hybrid inverter should I get", the answer is built from sites the AI could read. Cluttered HTML and blocked bots mean your store is not in that answer.

Bots are already here, and they are not free

In 14 days our store received 279 289 bot requests. More than half came from a single Meta training bot. That is server load you pay for, even without knowing it.

Reports do not show this

Standard analytics does not tell a ChatGPT visit from a direct one and does not show which bot reads what. Decisions get made blind.

Shopper asking ChatGPT what to buy while the assistant reads the store catalog

Everyone on the team gets their win

Store owner looking at a sales and AI traffic dashboard

«Shoppers already ask ChatGPT. If my store is not there, a competitor is.»

The pain:

  • A new sales channel showed up on its own, and the only cost so far is other people's training bots chewing the server I pay for.
  • Every new technology means a new contractor, new invoices, and new risk.

How RobotsLLM closes it:

  • The AI channel opens with one free add-on: no VPS, no contractor, no subscription.
  • Training bots are blocked by policy. In our case that is 45% of all bot requests, almost half of the bot load on the server.
  • The result shows up in money: 58 visits from AI assistants and 3 orders in the first 14 days, measured by the add-on, not by gut feel.
Store director checking that checkout still works late in the evening

«I measure every new thing with one question: what can it break on Friday night?»

The pain:

  • Modules that touch the core or robots.txt can take checkout down at the worst moment.
  • Easy to ship. Support and rollback usually belong to nobody.

How RobotsLLM closes it:

  • A regular add-on with no core patches: install publishes nothing and changes nothing on the storefront, uninstall removes everything it added.
  • Every robots.txt change goes through a before/after diff and rolls back to any version in one action.
  • A nightly self-audit checks files, headers, and content parity, so you see the problem before the fallout.
Head of SEO reading a report on bots and AI referrals

«Leadership asks about AI SEO, and I have neither the tools nor the numbers for a report.»

The pain:

  • Experiments with llms.txt and mirrors are scary to ship: you can nick Google indexing.
  • robots.txt is edited blind, and a mistake in it costs traffic.

How RobotsLLM closes it:

  • Mirrors and llms.txt are served with noindex, so they do not compete with HTML pages in Google.
  • An RFC 9309 compiler, a bot simulator, and a diff before publish: robots.txt is no longer a live edit.
  • A ready report for leadership: which bots came, IP verification, visits from ChatGPT, Perplexity, and Gemini all the way to orders.
Content manager working through a large product catalog

«Another content format? We have 4 600 products. I will not do this by hand.»

The pain:

  • Every new channel still meant rewriting cards by hand and weeks of spreadsheets.
  • Nobody has time to watch that nothing extra leaks into public files.

How RobotsLLM closes it:

  • Markdown versions of products, categories, and pages generate themselves from existing cards: in our store there are 18 000+ of them, and not one was written by hand.
  • You change a product: the queue updates its mirror in minutes.
  • The sanitizer strips scripts and trackers, and the secrets scanner blocks publish if something extra landed in the content.

One add-on instead of a server, scripts, and guesswork

AI analytics

Which bots visit, IP verification, visits from ChatGPT, Perplexity, Gemini, and orders from them, in the admin panel.

279 289bots
97visits
3orders

Smart robots.txt

Presets Growth, Open, Strict, or your own. Compiler per RFC 9309, bot simulator, "before → after" diff before publish, and rollback in one action.

diff
+ User-agent: GPTBot
+ Disallow: /
+ User-agent: ChatGPT-User
+ Allow: /

Registry of 31 bots

GPTBot, ClaudeBot, PerplexityBot, Bytespider and others, with each bot's purpose: search, training, user, SEO tools.

Content protection

A sanitizer strips scripts and trackers from mirrors; a secrets scanner blocks publish if something extra landed in the content.

Self-audit every night

The add-on checks its own files: HTTP statuses, headers, content parity. Problems show up before the fallout.

llms.txt

A store map for AI: categories, products, policies. Curation, language versions, token budget.

Markdown mirrors

Clean .md versions of categories, products, and pages. We have 18 000+ of them. Google does not index them: we serve noindex.

An llms.txt generator makes one file. RobotsLLM does the rest.

Capability llms.txt generators RobotsLLM
llms.txt with curation and languages
Markdown mirrors none 18 000+ on our store
Per-purpose bot policy none presets + RFC 9309
Diff before publish and rollback none
Registry of 31 bots with IP verification none
AI referrals down to orders none
Sanitizer and secrets scanner none
Nightly self-audit none
Price paid free in 1.0

Three steps, zero core patches

  1. 1

    Install the zip

    A regular CS-Cart add-on: upload, activate. No core patches, no separate VPS. Uninstall removes everything.

  2. 2

    Choose a policy and inspect the diff

    The wizard suggests Growth by default. You see the exact robots.txt diff before publish. Install itself changes nothing: you publish.

  3. 3

    Enable cron and collect numbers

    The generation queue, log parser, and nightly audit run through your host's regular cron. The dashboard fills itself.

Careful by design

  • Diff before every robots.txt publish
  • Rollback to any version in one action
  • The physical robots.txt takes priority until you decide otherwise
  • Mirrors with X-Robots-Tag noindex, so they do not compete with HTML in Google.

What we do not promise

Nobody can guarantee a ChatGPT citation or a percentage of AI-traffic growth. Anyone who promises that is selling air. RobotsLLM gives visibility, control, and honest numbers; you make the decisions.

robots.txt versions
$ robots.txt versions

  #3  2026-08-29 14:02  Growth
  #2  2026-08-22 09:11  Open
  #1  2026-08-16 18:40  Strict

→ rollback #2 … done

CS-Cart today. More to come.

The RobotsLLM core does not depend on the platform. The CS-Cart adapter is already in production, the next ones are in progress.

Tell me when my platform ships

Version 1.0 is free. Entirely.

$0

All 7 modules, no limits and no trials

  • robots policies with diff/rollback
  • llms.txt + mirrors
  • analytics
  • protection
  • audit
Download free

Everything in free 1.0 stays free. Pro will add deeper analytics and scale.

FAQ

Native llms.txt covers one file. RobotsLLM adds what core does not: per-bot access policy with diff and rollback, a verified bot registry, Markdown mirrors, a sanitizer and referral-to-order analytics. Keep the native llms.txt or replace it with a curated one. 4.19 and 4.20 compatibility is being verified; confirmed on 4.18.3.

No, and it will not harm them. Mirrors are served with noindex, and you see robots changes in a diff before publish. The add-on's goal is AI visibility and control; it carefully complements classic SEO.

Nobody can guarantee that. RobotsLLM removes technical barriers (unreadable HTML, blocked bots) and measures the result. The AI decides what to cite.

No. This is a regular add-on: PHP, a MySQL queue, and your host's cron. Files are served with ETag and 304, so bots do not load PHP for nothing.

Install does not change it at all. Publish shows a diff, every version is stored, rollback is one action. The physical file in the root always takes priority.

CS-Cart 4.18+ (PHP 8.1+) today. The core is platform-independent; OpenCart, Horoshop, and WordPress/WooCommerce adapters are in progress.

Version 1.0 is entirely free. Future Pro features will not take away what you got for free.

From the RobotsLLM dashboard on our store akumulyator.center for 16-29 August 2026. This is our own dogfooding.

See what AI sees in your store

Free. 10 minutes to install. Rollback always nearby.