• New
Markdown4Agents Pro – Markdown for AI in PrestaShop | PrestaDev
Markdown4Agents Pro – Markdown for AI in PrestaShop | PrestaDev
Markdown4Agents Pro – Markdown for AI in PrestaShop | PrestaDev
Markdown4Agents Pro – Markdown for AI in PrestaShop | PrestaDev
Markdown4Agents Pro – Markdown for AI in PrestaShop | PrestaDev
Markdown4Agents Pro – Markdown for AI in PrestaShop | PrestaDev
Markdown4Agents Pro – Markdown for AI in PrestaShop | PrestaDev
Markdown4Agents Pro – Markdown for AI in PrestaShop | PrestaDev

Markdown4Agents Pro for PrestaShop — store content in Markdown for AI agents

Module version: 1.2.0

PrestaShop compatibility: 1.7.x 8.x 9.x

Module file updated: 09/11/2026 17:52:14

Product data updated: 09/14/2026 09:41:22


Markdown4Agents Pro is a PrestaShop module that makes store content available to AI assistants and agents in Markdown format, instead of forcing them to process the full HTML page. The module works through two independent channels: content negotiation on the canonical URL of the supported page (a request with the Accept: text/markdown header receives Markdown, while the browser still receives HTML) and parallel .md URLs for agents that do not send the header. The document is generated directly from PrestaShop data — product, category, CMS page, manufacturer, and the store homepage itself — rather than by cleaning rendered HTML, so it contains prices, variants, features, and availability as data, not as text to be guessed. Optional integrations add blog posts from ph_simpleblog, questions and answers from FAQ modules, and customer ratings and reviews of products. The module also generates llms.txt and llms-full.txt files, i.e. a machine-readable table of contents for the store compliant with the increasingly widespread convention for language models. A separate Content Signals tab lets you declare in the robots.txt file what you allow automated systems to do with store content — indexing, answering customer questions, and training models — in accordance with the contentsignals.org specification. The built-in diagnostics act as a safeguard: content negotiation is installed disabled and cannot be enabled until the test confirms with real HTTP requests that this store’s stack — CDN, reverse proxy, cache — will not serve Markdown to the browser or HTML to the agent. The generated Markdown is cached in the database and invalidated by hooks when a product, combination, promotion, stock status, category, CMS page, or manufacturer changes, so it is not regenerated on every request. The module consistently enforces the rule that an agent never sees more than an unauthenticated guest: it respects quantity hiding, products unavailable for order, catalog mode, guest price hiding, product visibility, maintenance mode, and geolocation. The dashboard shows the number of handled requests, recognized agent and client families, and the most frequently fetched documents, including blog posts.

zł 278.00
zł 226.02no tax
Price history:

Description

Module features and benefits

  • Markdown built from PrestaShop data and supported content modules, not by converting the rendered page — the document does not contain menus, footer, forms, or code from other modules.
  • Radically lower token cost on the agent side: the x-markdown-tokens and X-Markdown-Tokens-Estimate headers in every response provide the estimated document size, and the panel shows a Markdown vs HTML comparison for the selected entity.
  • Two independent delivery channels — content negotiation at the canonical URL and separate .md URLs — so the agent reaches the content regardless of whether it sends the Accept header.
  • A diagnostic safeguard before enabling negotiation: the module does not allow the feature to be activated until it measures with real HTTP requests that the caching intermediary will not poison the shared cache.
  • The rule “the agent never sees more than a guest” implemented as a rule, not as isolated exceptions — store visibility settings apply to agents exactly as they do to an anonymous visitor.
  • Prices calculated in the guest context, with explicit currency and country resolution, so the cache will not store one customer’s price and serve it to everyone else.
  • Full multistore and multilingual support: settings, llms files, cache, and counter are maintained separately for each store and each language.
  • Zero external dependencies — no Composer, no external libraries, no calls to third-party services; the module does not send your store’s data anywhere beyond the response to the request.
  • Statistics without visitor identifiers: stored data includes daily aggregates, store, channel, client family, and the type, identifier, and language of the document with the time of the last request, without IP addresses, sessions, or the full User-Agent.
  • Everything agent-readiness validators look for: content negotiation on the main store URL, the Vary: Accept header, the token counter under the standard x-markdown-tokens name, and the Content Signals declaration in robots.txt — without manually adding anything to server files.
  • Compatibility from PrestaShop 1.7.1.0 to 9.x and from PHP 7.0 to 8.5, verified against real source trees rather than declared only.

Key module functionalities

  • Content negotiation on canonical store URLs: a request with Accept: text/markdown receives a Markdown document, while a browser request receives unchanged HTML.
  • The Vary: Accept header and the Link: rel="alternate" tag on HTML pages covered by the module, so intermediaries and agents know that an alternative version exists.
  • Parallel markdown/{typ}/{id}.md URLs for products, categories, CMS pages, manufacturers, the store homepage, and integrated blog articles, under its own prefix so they do not conflict with other URL-rewriting modules.
  • The Integrations tab: support for ph_simpleblog, pdfaqspro, pdproductsfaqpro, iqitreviews, and productcomments, with separate selection of the scope of published content.
  • Statistics dashboard: period, channel, and client family filters, daily chart, agent and client ranking, and the 20 most frequently fetched documents.
  • The Content Signals tab: declaration of preferences toward automated systems, published in robots.txt — details in a separate section below.
  • The store homepage also responds in Markdown — it gets its own document with the store name and description, main categories, and informational pages. This is the URL checked by agent-readiness validators by default (isitagentready.com, the Agent Diagnostics panel in Cloudflare), so without it, a store with perfectly working negotiation across the entire catalog was still marked as not ready.
  • The llms.txt file — a short list of the store’s most important sections: top-level categories, informational pages, selected blog articles, and a link to the full index.
  • The llms-full.txt file — a compact index of public entities, including blog articles. It contains titles, links, and summaries, and for products also available offer data plus the average rating and number of reviews from the selected review source. Automatically split into parts after exceeding the configured size.
  • Generation of llms-full.txt via a token-protected cron endpoint or manual rebuild in the panel, with a time budget and interrupted job resumption. The full index is not built during a normal visitor request.
  • A separate cron token for each store, compared in constant time, with a ready-to-copy curl command in the panel.
  • Cache of generated Markdown in the database with configurable lifetime and hook-based invalidation when a product, combination, promotional price, price rule, stock status, category, CMS page, or manufacturer changes.
  • The Diagnostics tab: seven checks performed with real HTTP requests against the module’s own probing endpoint, with separate distinction between failures of “HTML served to the agent” and “Markdown served to the browser”.
  • An informational check of the real catalog URL after negotiation is enabled, presented separately and not gating the feature.
  • Document preview in the panel for the selected entity type and identifier, together with a Markdown vs HTML token counter and an explicit entity visibility verdict.
  • The Files tab: status of llms files for each language (presence, size, date, number of entities, number of parts), a manual rebuild button with a time budget, and the cron command.
  • Agent hit counter: aggregate day × store × channel × client family, split into negotiation, .md route, and llms channels, and openai, anthropic, perplexity, google, bing, script, browser, and other families.
  • Response headers aligned with the content purpose: X-Robots-Tag: noindex, nofollow on Markdown documents, X-Content-Type-Options: nosniff, ETag with 304 support, and a differentiated Cache-Control policy for the canonical URL and the .md URL.
  • Product data in the document header: type, identifier, canonical URL, Markdown URL, title, meta description, main image URL, SKU, brand, gross and net price, currency, tax information, availability, pack designation, personalization requirement, product condition, categories, language, and update date.
  • Regular price as well as the discount amount and percentage when the product is on promotion — the agent will not see the promotional price as if it were the regular one.
  • A variants table with price, quantity, availability, and combination identifier, so the agent can point to a specific variant instead of only describing it.
  • Contents of product packs with quantities and links to the Markdown documents of individual items.
  • Information about required personalization fields, excluding fields belonging to other modules that PrestaShop does not show on the frontend.
  • Selection of entity types made available to agents, enforced consistently across all channels: on .md routes, in llms files, and in the alternate-version tag.
  • The module’s working directory secured against direct web access, and llms files served exclusively by module controllers.
  • Respect for store maintenance mode and geolocation — a store closed to visitors is also closed to agents.
  • A panel in the PD convention with the module header and six tabs: Dashboard, Settings, Integrations, Diagnostics, Files, and Content Signals. The Dashboard is the start screen, and the settings are accompanied by explanations of how they work.

Blog and guides — integration with ph_simpleblog

The integration also makes the store blog’s knowledge available to agents: buying guides, instructions, and informational articles. Each published article available to a guest can have its own Markdown document and a link in the llms files.

  • The article document contains the title, introduction, content, author, blog category, publication and update dates, and the source page URL.
  • The activity of the entry and its category, publication date, guest access, store assignment, and translation availability are respected. Private articles and those scheduled for the future are skipped.
  • Public articles go into llms-full.txt, and the short llms.txt includes a blog and guides section with titles, links, and available short descriptions.
  • From 0 to 20 articles can be displayed in the short overview. A custom list of identifiers determines the order; when the list is empty, the module selects featured articles first and then the newest ones. Setting 0 hides this section while keeping the articles in the full index.
  • Articles use the same access channels as the catalog: the Markdown URL and content negotiation on the source page when the relevant features are enabled. Their documents can be checked in the panel preview.

Product, category, and CMS page FAQs

Questions and answers form a section of the document for the page they relate to. The agent receives the offer description or page content together with answers to related questions, for example about installation, product selection, usage, or service terms.

  • pdfaqspro provides FAQs assigned to products, categories, and CMS pages. Each of these scopes can be enabled separately in the Integrations tab.
  • The older pdproductsfaqpro is supported as a source of product FAQs. Category and CMS FAQs require the newer pdfaqspro. If both modules are active, the newer variant takes precedence.
  • Published questions are active ones from the correct store and the correct page, taking into account publication settings and the language rules of the source module. If the source allows a fallback language, the document indicates the language of the attached answer.
  • One document can contain up to 500 questions and answers, in the order defined in the FAQ module. Empty questions or answers are skipped.
  • The llms files lead to product, category, and CMS documents with their FAQ sections. Individual questions do not create additional index entries.

Product ratings and reviews — iqitreviews and productcomments

The product document may contain a summary of customer ratings and the content of the latest public reviews. In the panel, you choose one source: iqitreviews or productcomments, so reviews from two modules are not combined.

  • The average rating and number of reviews are calculated based on all public reviews according to the rules of the selected module, taking moderation and hidden entries into account.
  • Two scopes are available: average rating and number of reviews only, or the summary together with 1–20 latest statements. The text limit does not restrict the set of ratings used to calculate the average.
  • Statements are selected from newest to oldest, without filtering by rating value. They retain their original language and are clearly marked as customer reviews.
  • The data scope of a single review includes rating, date, title, and content. The module does not attach fields with the customer name or account identifier.
  • In llms-full.txt, a product may receive a concise note about the average rating, number of reviews, and source. The review texts remain in the product document.

The Integrations tab — scope and freshness of content

The panel shows which source modules are active in the selected store. You can separately enable the blog, product FAQs, category FAQs, and CMS FAQs, and choose the review source and scope. Integrations are optional, disabled by default on installation, and their settings apply to a specific store.

Changes and deletions of articles, FAQs, and reviews are checked when a Markdown document request reaches the store, including when the document is already in the module cache. The llms.txt and llms-full.txt files are refreshed during rebuild in the Files tab or via CRON. Copies stored by a CDN and other external caches are subject to the configured response lifetime.

Dashboard — agent statistics and most frequently fetched content

The Dashboard lets you see which parts of the store are fetched in the formats provided by the module. It includes catalog documents, blog articles, and llms files.

  • Selection of the last 7, 14, or 30 days and filtering by access channel and agent or client family.
  • Summary of the number of requests, the share of recognized bots and AI agents, the number of different documents, and requests recorded with content information.
  • Daily chart with a value table, ranking of agent and client families, and channel comparison: content negotiation, Markdown URLs, and llms files.
  • Ranking of the 20 most frequently fetched documents with title, type, language, number of requests, share in content traffic, and time of the last request. Blog articles appear with their own title and link.
  • Statistics for the selected store in multistore. Content details are collected from the deployment of the extended counter; earlier request totals remain available.

The counter measures requests handled by the module, including conditional 304 responses, not unique visitors. Client family recognition is based on the declared User-Agent. Requests handled entirely by external cache do not enter the module statistics.

Content Signals — a declaration of what you allow automated systems to do

A store whose content is read by AI agents will sooner or later face the question of what it actually agrees to. Is it allowed to build a search engine index from it? Is it allowed to use a product description to answer a customer’s question? Is it allowed to train a model on this catalog? Content Signals is a standard that lets you answer these three questions separately and publish the answer where automated systems already look anyway — in the robots.txt file. The module provides a separate tab for this, three simple choices, and a preview of exactly what will go into the file. Specification: contentsignals.org.

  • Search engine indexing (search) — building an index and showing links with short snippets. This does not include AI-generated summaries.
  • Answering questions based on your content (ai-input) — providing the content to a model at the moment someone asks a question so the answer is based on it. This is exactly how shopping assistants work.
  • Training AI models (ai-train) — using the content to train or fine-tune models.
  • Each of the three categories has an independent setting of allowed or not allowed, described in the panel with one sentence in Polish — not just a technical label that a merchant is not expected to know.
  • The declaration goes into robots.txt in a block delimited by markers with the module name. Nothing outside those markers is parsed, rewritten, or rearranged — your own rules and the rules of other modules remain untouched byte for byte.
  • PrestaShop overwrites robots.txt from scratch every time you click “Generate robots.txt file” in SEO and URLs. The module then automatically writes its block back, and the tab still shows whether the file is up to date and lets you save it again with one button.
  • Before saving, the panel shows whether the file is writable at all — instead of reporting success and leaving you with a declaration that is not actually in the file.
  • The feature is installed disabled: writing anything into a file shared by the entire store is the merchant’s decision, not the module’s. Uninstalling the module removes the block again.
  • In a multistore installation, the panel explicitly warns that PrestaShop keeps one robots.txt file for all stores, so the saved declaration covers every one of them.
  • This is a declaration of preference, not access control: it does not block anyone, and whether it is respected is the other party’s decision. The module says this explicitly instead of suggesting protection that this mechanism does not provide.

Business use

  • For PrestaShop stores that want to be correctly read by shopping assistants and AI agents, instead of relying on the model to clean up the HTML page on its own.
  • For merchants who want to control exactly what reaches agents — the scope of entity types, the visibility of prices and quantities, and the publication method.
  • For stores with an extensive catalog, where the difference between a full HTML page and a Markdown document translates into a real cost difference on the querying side.
  • For stores running a blog and FAQ base that want to make guides, answers to questions, and customer reviews available to agents together with the product offer.
  • For multistore deployments, where each store requires its own settings, its own llms files, and its own cron token.
  • For companies that want to base decisions about further investment in this channel on data — the hit counter shows whether agents are actually querying and through which path.
  • For stores operating behind a CDN or reverse proxy, where enabling content negotiation on your own without verification could risk serving the wrong page version to regular customers.

What the module does not do

  • It is not an SEO module and does not improve the store’s position in Google search results — Markdown documents are explicitly marked as non-indexable.
  • It preserves the appearance and content of the HTML page seen by customers, adding information about the alternative Markdown version in HTTP headers and the link tag.
  • It does not send store data to any external service and does not require an account or API key.
  • It does not enforce Content Signals and cannot do so: the declaration in robots.txt is an expression of preference, not access control — it does not block anyone, and whether it is respected remains the other party’s decision. The module does not modify the remaining rules in that file.
  • It does not perform automatic translations. Document field names remain fixed and in English, page and article content uses available translations, FAQs respect the language rules of the source module, and reviews retain the language of the customer’s statement.

Compatibility

  • PrestaShop: 1.7.1.0 - 9.x
  • PHP: 7.0 - 8.5
  • No Composer and no external libraries
  • Multistore: yes, settings and files maintained separately for each store
  • Multilingual: yes, documents and llms files for each active language

Product Details

Prestashop 9.x support
Yes
Support for Prestashop 8.x
Yes
Prestashop 1.7.x support
Yes
Module translations
EN, ENG
Free support
Yes
Free updates (1 year)
Yes
Easy Installation
Yes
PDMD4APRO

Changelog

Release history for this product. Expand an entry to see what changed.

  • dodano obsługę Content Signals
  • dodano obsługę ph_simpleblog
  • dodano obsługę pdfaqpro
  • dodano obsługę productcomments
  • dodano obsługę iqitrewievs

FAQ

Find answers to the most common questions about this product.

The module makes PrestaShop store content available in Markdown format, intended for AI assistants and agents. It operates through two channels: at the canonical product URL, it responds in Markdown when the request includes the Accept: text/markdown header, and for clients that do not send such a header, it provides parallel URLs ending with the .md extension. Additionally, it generates llms.txt and llms-full.txt files, which serve as a machine-readable table of contents for the store. The supported content types are products, categories, CMS pages, and manufacturers.

No. This is not an SEO module and it does not affect rankings in search results. Markdown documents are explicitly marked with the X-Robots-Tag: noindex, nofollow header, which directly asks search engines not to index them. The module addresses a different question than SEO: how to ensure that an AI assistant reads the store’s offer correctly and cost-effectively, instead of processing the entire HTML page together with the menu, footer, and the code of other modules.

The module does not convert the rendered page; instead, it builds the document directly from PrestaShop data. As a result, the price, regular price during a promotion, currency, tax information, availability, SKU, features, variants with quantities and identifiers, and bundle contents are included in the document as structured data rather than text that has to be guessed. HTML conversion, on the other hand, provides whatever happens to be present in the presentation layer, along with interface elements and content from other modules.

No. The customer's browser still receives a regular HTML page, with no changes to its appearance or content. Pages covered by the module only receive additional headers informing intermediaries and user agents that an alternative version exists: Vary: Accept and Link rel="alternate". The module does not modify the theme, does not add scripts to the front end, and does not change the way the page is rendered.

These are two text files in the convention adopted by tools based on language models. llms.txt is a short, curated list of the store’s most important sections: top-level categories, information pages, and a link to the full index. llms-full.txt is a compact index of entities, where each entry contains a title, canonical URL, Markdown version URL, price, availability, SKU, and a short summary, while the agent retrieves the full content only from the document it is interested in. The file is automatically split into parts once the configured size is exceeded.

This is an intentional safeguard. Content negotiation means that under a single address, the response depends on the request header, and this is dangerous if any cache along the way — CDN, reverse proxy, hosting cache — stores one version and serves it to everyone. Then an ordinary customer could see raw text instead of the store page. That is why the switch is locked until the built-in diagnostics verify with real HTTP requests that your store stack behaves correctly in both directions. The diagnostics themselves do not change any settings.

For full operation, yes, but not for running. The llms-full.txt file is built exclusively by a token-protected endpoint and is never generated during a visitor's request, so that a large catalogue does not block the page. In the admin panel, there is a manual rebuild button and a ready command to copy into the task scheduler. Without cron, the module works, and the llms-full.txt files are refreshed only when you click rebuild manually. Each shop has its own separate token.

No. The module consistently follows the principle that the agent never sees more than an unauthenticated guest on the store page. The following are respected: hiding stock quantities, products marked as unavailable for order, catalog mode, hiding prices from the Guest group, product visibility settings, as well as store maintenance mode and geolocation. Prices are calculated in the context of an anonymous visitor, so the cache will not save one customer's price and show it to others.

No. The module does not call any external service, does not require an account or an API key, and does not send any store data beyond the response to an incoming request. The built-in hit counter stores only aggregated data: the day, store identifier, channel, and customer family reduced to a few categories. It does not store IP addresses, session identifiers, full URLs, or the full contents of the User-Agent header.

This is handled by the built-in hit counter. It shows the number of requests broken down by day, shop, channel, and the requesting family. Channels include content negotiation, .md addresses, and llms files, while families include, among others, openai, anthropic, perplexity, google, and bing, as well as requests from scripting libraries and browsers. This means the decision to further develop this channel is based on data from your shop, not on assumptions.

Yes, in both cases fully. Settings, llms files, cache, cron token, and hit counter are maintained separately for each store. Markdown documents and llms files are generated for each active store language. Saving settings requires selecting a specific store in the switcher so that one click does not overwrite the configuration of the other stores.

The module works with PrestaShop from 1.7.1.0 to 9.x and PHP from 7.0 to 8.5, without Composer and without external libraries. The generated Markdown is stored in the database cache and invalidated automatically when the product, combination, promotion, stock status, category, CMS page, or manufacturer changes, so it is not regenerated on every request. The HTML page seen by customers is neither rebuilt nor slowed down — it only receives additional headers.

Reviews (0)

No reviews