PrestaShop Agentic commerce & AI agents

Our best PrestaShop modules for agentic commerce and AI agents

Six modules that make the store readable, measurable and buyable by AI agents: MCP, ACP, llms.txt, crawler control and radar.

AI assistants no longer just recommend, they are starting to buy. Agentic commerce makes the store readable and operable by agents: from the cited catalogue to the order created in a chat and landing in your back office.

The problem

Sound familiar?

Invisible to the assistants

ChatGPT and Perplexity answer product questions: if your catalogue is unreadable for models, they recommend the competitor.

The recommendation without the transaction

The customer wants to order in the chat, but the path ends at a link: the agent can recommend, not buy.

Bots without rules

AI crawlers read everything or nothing: without control you either give away too much or vanish from the answers.

The invisible channel

AI traffic barely shows in classic analytics: the new channel grows and nobody sees it.

The shortlist

Our selection, ranked

Every module below is built, maintained and supported by our team. The ranking reflects what we would install first on a client store.

  1. MCP server for PrestaShop: catalogue, stock and orders as tools any MCP-capable AI assistant can call.

    Expose your PrestaShop store to AI agents (ChatGPT, Claude) through an MCP server. Catalogue, cart and order driven by AI, with OAuth 2.1 and…

  2. Agentic Commerce Protocol (ACP): ChatGPT builds the cart and triggers the order, which lands in your normal back office.

    Expose your catalogue and checkout to AI agents (ChatGPT, Claude, Perplexity) through the Agentic Commerce Protocol. 5 conformant REST endpoints, real PrestaShop order, Stripe…

  3. llms.txt and llms-full.txt generated automatically: language models read your catalogue in a structured way instead of guessing.

    Expose your PrestaShop catalogue to LLMs (ChatGPT, Claude, Perplexity, Gemini) via the standard llms.txt and llms-full.txt files. Automatic generation, smart cache, multistore & multilingual.

  4. AI crawler manager: rules per bot, who may index and what they see. The channel stays open, the control stays yours.

    Fine-grained AI bot management (GPTBot, ClaudeBot, PerplexityBot, Google-Extended, Bytespider and 25+ more) with visual robots.txt builder, HTTP 403 blocking and crawl statistics. PrestaShop 8…

  5. Real-time traffic distinguishing humans from AI crawlers: the new channel becomes visible and measurable.

    Turn your back office into a traffic control tower: live visitors, a clean humans / bots / AI crawlers split, charts, heatmap and AI…

  6. Webhooks to Zapier, Make and n8n: orders and events flow into your automations, wherever they come from.

    Bidirectional no-code webhook connector: link PrestaShop to Zapier, Make, n8n and 5000+ apps, with an async queue, automatic retries and HMAC signing.

Side-by-side comparison

The next customer is an agent

ChatGPT, Claude and Perplexity no longer just recommend products: they are starting to buy them. Agentic commerce means an AI assistant searches, compares and orders on the customer’s behalf. A store an agent cannot read and cannot operate does not exist for that channel.

Become readable, then buyable

llms.txt makes the catalogue readable for language models. The crawler manager controls which AI bots see what. The MCP server exposes catalogue, stock and orders as tools an agent can call. And the Agentic Commerce Protocol lets ChatGPT pay directly in your store.

And see the new traffic

AI traffic barely shows in classic analytics. The real-time radar tells humans and AI crawlers apart, and webhooks connect the orders to the rest of your automation.

Buying guide

How to choose

Readable first, buyable second

Start with llms.txt and the crawler manager: the catalogue becomes readable for models and you control who sees what. That is the foundation recommendations grow on. Only then is the transaction layer worth it.

MCP for depth, ACP for reach

The MCP server makes the store operable by any MCP-capable assistant: maximum flexibility, including your own agents and B2B integrations. ACP targets the ChatGPT checkout: a single but enormous channel. The two complement each other.

Treat agents as a sales channel

An agent compares in a structured way: price, availability, delivery time, return conditions. Stores with clean machine-readable data win those comparisons. The data quality Google Shopping demanded is now demanded by every assistant.

Measure before you scale

The radar shows which AI crawlers come and how often. Watch the channel for a few weeks before setting priorities: the distribution between assistants changes fast, and your data says more than any forecast.

What you gain

The store becomes an agent interface

Catalogue, stock, prices and orders as MCP tools: an AI agent can query and operate your store like an API client.

The sale inside the chat

Checkout via ChatGPT through the Agentic Commerce Protocol: the order is created in the conversation and lands in your back office.

A catalogue models can read

llms.txt and llms-full.txt describe the catalogue in a structured way: language models cite your products instead of guessing them.

AI bots under control

Rules per AI crawler: who may index, who may not, and what they see. Control instead of blind flight.

Measure and connect

Real-time radar for humans and AI crawlers plus webhooks to Zapier, Make and n8n: the new channel becomes visible and automatable.

Implementation

From install to results

  1. Make the catalogue readable

    Generate llms.txt and set crawler rules: the catalogue becomes readable, the control stays.

  2. Measure the AI traffic

    Run the radar for a few weeks: see which assistants come.

  3. Open the agent interface

    Enable the MCP server: catalogue and orders become operable by agents.

  4. Enable the ChatGPT checkout

    Implement ACP: the order is created in the chat and lands in the back office.

  5. Wire up the rest

    Webhooks to your automations: the new channel fits into existing processes.

“We installed llms.txt and the radar to see whether the topic was real. Three months later, measurable orders come through assistant recommendations. The channel exists, and it is growing.”

Customer feedback — PrestaShop 9 store, technical accessories

Frequently asked questions

What separates agentic commerce from AEO?

AEO (Answer Engine Optimization) makes AI assistants know and cite your products: visibility. Agentic commerce goes a step further: the agent can query the catalogue, build the cart and trigger the order. One brings the recommendation, the other the transaction.

What is an MCP server concretely?

MCP (Model Context Protocol) is the open standard AI assistants use to call external tools. The MCP server exposes your PrestaShop as such tools: product search, stock query, cart, order. Any MCP-capable assistant can use them, with the permissions you grant.

How does the ChatGPT checkout (ACP) work?

The Agentic Commerce Protocol is the standard initiated by OpenAI through which ChatGPT triggers orders directly with merchants. The module implements the endpoints on the PrestaShop side: product feed, checkout session and payment handling, with the order landing in your normal back office.

Should I not just block AI crawlers?

No, that is a misjudgement. AI crawlers decide whether your products appear in answers. The crawler manager allows a differentiated policy: catalogue yes, prices delayed, internal pages no. Blocking wholesale means giving up the channel.

Is this compatible with PrestaShop 9?

Yes. The whole selection covers PrestaShop 8.0 to 9.x, using native hooks and no core overrides.

This need on other platforms

Not sure which one fits your store?

Tell us your context — we answer with a straight recommendation, not a sales pitch.