MxChat WooCommerce

An AI shopping assistant for WooCommerce stores. Customers talk naturally with the chatbot — it renders product cards inside the chat, recommends complementary items, adds to cart, checks order status, and runs intelligent filtered search, all in conversation.

Overview

MxChat WooCommerce turns your store's chatbot into a product-aware sales assistant. Instead of pointing customers at search results or category pages, the AI surfaces the right products directly in chat with real images, real prices, and a working Add to Cart button. It understands contextual references like "add it" or "the red one", handles price-range and attribute filters in natural language, and pulls real-time order history for logged-in shoppers. The add-on integrates with WooCommerce's product catalog, cart, and orders APIs, and is scoped per bot so multi-store networks stay clean.

Setup

Install the plugin via Plugins → Add New → Upload Plugin, or search the WordPress repo for "MxChat WooCommerce".
Activate via Plugins → Installed Plugins. The add-on will warn you if MxChat (mxchat-basic) or WooCommerce is missing.
Make sure your MxChat Pro license is active under MxChat → Activation. WooCommerce features are Pro-only.
Open MxChat → WooCommerce to configure product display, recommendation, and cart settings.
Index your catalog: enable Auto Product Display so products are surfaced automatically when the AI mentions them, and enable the relevant intents under MxChat → Intents (Product Card, Add to Cart, Order History, Recommendations, Checkout, Filtered Search, Featured Products).
To show the chat cart and order icons, enable the toolbar under MxChat → Toolbar & Components.

Settings

Auto Product Display

Controls when product cards appear automatically below an AI response. Dropdown with three values:
- Off — never auto-render product cards.
- All Matched Sources — show a card for every product the RAG retrieval matched (recommended for most stores).
- Only AI-Mentioned Products — show cards only for products the AI actually referenced in its reply.

Product Display Options

Toggles for the product card layout: show prices, show Add to Cart button, custom Add to Cart Button Text, and custom View Details Button Text. Empty values fall back to the translated defaults.

Recommendation Strategy

The Product Recommendations intent queries WooCommerce directly across titles, descriptions, and SKUs, then falls back to featured / bestsellers / newest when results are thin. Displays up to 4 recommendations with images, pricing, and the AI's reasoning.

Filtered Product Search

Configurable trigger phrases (e.g. "under $", "less than", "cheaper than"), adjustable similarity threshold for trigger detection, and an optional semantic re-ranking toggle for better relevance. Out-of-stock products are excluded automatically and manual price validation guarantees only correctly-priced products appear.

Featured Products Showcase

An admin product picker with live search lets you curate a hand-picked set of products. An AI context prompt input controls how the chatbot frames them when presenting the showcase.

Cart Handling Rules

The Add to Cart intent uses AI-powered product resolution: it analyzes recent conversation history to determine which product the customer means, supports contextual references in any language, and gracefully falls back to multi-term WooCommerce search if AI resolution is unavailable. Catalog lookup is capped at 100 products for performance.

Features

Product Card Intent: Renders a product card in the chat with image, name, price, and a working Add to Cart button — no page reload required.
Add to Cart Intent: AI identifies the right product from natural language ("add it", "that one", "the second one") and drops it in the WooCommerce cart. Works in any language and handles variable products with attribute selection.
Intelligent Filtered Product Search: Natural-language product discovery — "show me products under $30", "red shoes in stock", "laptops between $500 and $1000". Extracts price ranges, categories, attributes, and search terms automatically and shows up to 6 product cards.
Product Recommendations: Up to 4 AI-curated suggestions based on the customer's order history, current cart, and ongoing chat. Each card includes image, price, and the AI's reasoning.
Order History Intent: Logged-in customers ask about past orders by date, product, or status and get instant answers without leaving chat.
Checkout Intent: Guides customers from product discovery straight to the WooCommerce checkout page.
Featured Products Showcase: Curated product picker with an AI-generated intro message — great for highlighting sales, new arrivals, or staff favorites.
Auto Product Display: When the AI mentions WooCommerce product URLs, matching product cards render automatically below the reply. Works with both streaming and non-streaming responses.
Multilingual support: All AI responses detect and match the customer's language. Filtered search even handles cross-language queries (Spanish customer, English catalog).
Theme-aware styling: Product cards pick up your chatbot's light or dark theme automatically and respect your display settings (show prices, show Add to Cart, custom button text).

API & Integration

The add-on registers seven WooCommerce-specific callbacks via the mxchat_available_callbacks filter:

mxchat_handle_product_inquiry — show a product card in chat.
mxchat_handle_product_recommendations — surface AI-curated recommendations.
mxchat_handle_filtered_search — natural-language filtered product search.
mxchat_handle_featured_products — render the curated featured-products showcase.
mxchat_handle_add_to_cart_intent — add a product to the WooCommerce cart from chat.
mxchat_handle_checkout_intent — redirect to the checkout page.
mxchat_handle_order_history — return logged-in customer order history.

It consumes core WooCommerce hooks for live catalog awareness: woocommerce_update_product and woocommerce_delete_product for product changes, plus wp_insert_post / wp_trash_post / before_delete_post when product sync is enabled. The mxchat_similarity_results action is used to pre-build product cards immediately after RAG retrieval so Auto Product Display renders instantly.

AJAX endpoints (all nonced via mxchat_nonce) are exposed for cart operations: mxchat_woo_get_cart_count, mxchat_woo_get_cart_contents, mxchat_woo_remove_cart_item, mxchat_woo_add_to_cart, plus product-card rendering endpoints (mxchat_woo_auto_product_cards, mxchat_woo_source_product_cards, mxchat_woo_prebuilt_product_cards). Variable products use mxchat_get_variation_data and mxchat_add_variation_to_cart.

Settings are scoped under the mxchat_woo_options option, so each bot_id / store in a multi-site install keeps its own configuration.

Requirements

MxChat (mxchat-basic): Required. Install and activate before this add-on.
MxChat Pro License: Required. WooCommerce features are Pro-only — see pricing.
WooCommerce: Required, version 6.0 or newer. Declares compatibility with HPOS (custom order tables).
WordPress: 5.8 or newer.
PHP: 7.4 or newer.

Troubleshooting

Products not appearing in chat

Check that the relevant intent is enabled under MxChat → Intents (Product Card, Filtered Search, Recommendations, etc.). If you expect Auto Product Display to render cards, confirm MxChat → WooCommerce → Product Display → Auto Product Display is set to "All Matched Sources" or "Only AI-Mentioned Products" — not "Off". For Filtered Search specifically, verify trigger phrases match your store's language and that the similarity threshold isn't set too high.

Inventory out-of-sync or stale prices

The add-on uses live WooCommerce calls, so stale data usually points at object cache. Clear your WordPress object cache and any page cache. If you're on managed hosting (e.g. WP Engine, Kinsta), exclude the /wp-admin/admin-ajax.php endpoint from full-page caching. Out-of-stock products are excluded from filtered search results automatically.

Recommendation quality is off

The Product Recommendations intent now queries WooCommerce directly across titles, descriptions, and SKUs (rewritten in v1.7.0). If results are thin, check that your products have rich titles and descriptions — the AI matches on real catalog text, not just embeddings. Featured / bestsellers / newest are used as automatic fallbacks when search returns too few products.

"API key is not set" error with Claude, Gemini, Grok, or DeepSeek

Fixed in v1.3.1 — make sure you're on 1.3.1 or later. The filtered search handler now respects the selected model's API key in core MxChat settings.

Cart and order icons require the MxChat toolbar to be enabled under MxChat → Toolbar & Components. The add-on injects its buttons into that toolbar; if the toolbar is off, the icons can't render.

Last reviewed by Sage on 2026-05-21. Spotted an issue? Open a ticket and we'll patch the doc.