Tutorial

Blazing-fast product search w. MeiliSearch and Medusa

Add blazingly fast product search to your e-commerce platform with the MeiliSearch plugin for Medusa.Post thumbnail image

Introduction

Search functionality is one of the most useful and important features in e-commerce platforms. From increasing customer conversion rates to significantly improving the user experience, search engines can enable significant business growth. Medusa brings search functionality to your doorstep by leveraging some of the already existing search engines out there.

We have developed a plugin that will allow you to use the performant, open-source, and feature-rich search engine MeiliSearch.

MeiliSearch is a super-fast, open-source, search engine built in Rust. It comes with a wide range of features such as typo-tolerance, filtering, sorting, and much more. MeiliSearch also provides a pleasant developer experience, as it is extremely intuitive and newcomer-friendly - so even if you're new to the search engine "ecosystem", you'll have a great time navigating through their documentation.

Through Medusa flexible plugin system, it is possible to enable search functionality into your medusa applications with minimum hassle by including our new plugin medusa-plugin-meilisearch to your medusa-config.js file.

Installation

In case you don't have MeiliSearch installed locally on your environment yet, you can run the following:

# Install MeiliSearch
curl -L https://install.meilisearch.com | sh

# Launch MeiliSearch
./meilisearch

For other installation alternatives, you can head over to Meilisearch's installation guide.

In order to add the plugin to your medusa project, you will need to first install the plugin using your favorite package manager:

# yarn
yarn add medusa-plugin-meilisearch

# npm
npm install medusa-plugin-meilisearch

Then in your medusa-config.js file, add the plugin to your plugins array. For this example, assumption is that you're leveraging MeiliSearch to perform searches on an index called products

module.exports = {
    // ...other options
    plugins: [
        // ...other plugins
        {
            resolve: `medusa-plugin-meilisearch`,
            options: {
                // config object passed when creating an instance of the MeiliSearch client
                config: {
                    host: "http://127.0.0.1:7700",
                    apiKey: "your_master_key"
                },
                settings: {
                    // index name
                    products: {
                        // MeiliSearch's setting options to be set on a particular index
                        searchableAttributes: ["title", "description", "variant_sku"],
                        displayedAttributes: ["title", "description", "variant_sku"],
                    },
                },
            },
        },
    ],
};

Et voilà! That's all it takes to make medusa and MeiliSearch work in great harmony. Please note that you can use any other setting from this list such as filterableAttributes, sortableAttributes, and synonyms. We are working on another blog post that will demonstrate how you can make use of these settings to build a great storefront experience, so stay tuned to that!

Simple Usage

Medusa exposes an API layer that can serve as an abstraction over various search providers. We will now be interacting with one of the search routes parts of that layer, namely, the POST /store/products/search route. The route will enable you to test out the integration between Medusa and MeiliSearch. The endpoint takes a mandatory q property and a set of optional parameters which will be passed to MeiliSearch's search() method as a second argument. The list of the parameters which can be provided can be found in MeiliSearch's docs.

Let's use Postman for this short demo to get some search results:

Postman screenshot showing search request and response

We tried to perform a search query "creneck" to find all crewnecks in our store. Note how the explicit mistake of leaving a "w" out from "crewneck" still yields the correct results.

Postman is not the most exciting client to use to showcase this, so in the next section, there is a short guide on how you can use React and MeiliSearchh's ecosystem to build something that looks better than JSON.

Highlighting and Pagination in a React Storefront

The Medusa + MeiliSearch integration opens up a lot of capabilities for creating a rich UX for your storefront. The plugin indexes all your products and listens to any updates made on them, so you can then leverage any client-side SDKs developed by the MeiliSearch team to build cool search experiences for storefronts. For example, MeiliSearch exposes a React adapter that can be used with React InstantSearch (made by Algolia) which provides features such as highlighting, filtering, and pagination out of the box.

In order to leverage this functionality, you'll need to install the corresponding packages by running:

# npm
npm install react-instantsearch-dom @meilisearch/instant-meilisearch
# yarn
yarn add react-instantsearch-dom @meilisearch/instant-meilisearch

You can then use the MeiliSearch client in your react app:

import React from 'react';
import {
  InstantSearch,
  Hits,
  SearchBox,
  Pagination,
  Highlight,
} from "react-instantsearch-dom";
import { instantMeiliSearch } from '@meilisearch/instant-meilisearch';

const searchClient = instantMeiliSearch(
  "http://127.0.0.1:7700",
  "your_master_key"
);

const SearchPage = () => (
  <InstantSearch
    indexName="products"
    searchClient={searchClient}
  >
    <SearchBox />
    <Hits hitComponent={Hit} />
    <Pagination />
  </InstantSearch>
);

const Hit = ({ hit }) => {
    return (
      <div key={hit.id}>
        <div className="hit-name">
          <Highlight attribute="name" hit={hit} />
        </div>
        <img src={hit.image} align="left" alt={hit.name} />
        <div className="hit-description">
          <Snippet attribute="description" hit={hit} />
        </div>
        <div className="hit-info">price: {hit.price}</div>
        <div className="hit-info">release date: {hit.releaseDate}</div>
      </div>
    )
}

If you want to play around with the various features provided by React InstantSearch you can read more on algolia's documentation. You can also use MeiliSearch's react demo for a more interactive example.

Demo: Palmes Storefront

By using the above libraries (React, react-instantsearch-dom, and instant-meilisearch) in addition to the medusa plugin, we have successfully integrated the MeiliSearch plugin for one of our customers: Palmes. Following is a short GIF demoing the functionality.

{% vimeo 639184340 %}

Enhance your development experience with MeiliSearch's Web UI

For a quicker feedback loop on what's happening on your search engine, you can use MeiliSearch's helpful out-of-the-box web interface to run some searches and get live results. It comes with MeiliSearch when you first install it and requires 0 configuration. When integrating the MeiliSearch plugin to existing medusa projects, we also found it to be extremely helpful for debugging

{% vimeo 639186298 %}

More to come

As briefly mentioned before, we're preparing another blog post that dives a bit deeper into the Medusa Search API and provides a more in-depth walkthrough on how to build a feature-rich search experience that includes filtering, synonyms, stop-words, and more!

Person photo
Zakaria El AsriOct. 26, 20214 min.

Want to know more about Medusa

Find related blog posts below

Tutorial

How to Create Custom Endpoints in Medusa

In this tutorial, you’ll learn how to create custom endpoints on your Medusa server for both the storefront and admin.

Tutorial

How to Create a Comic Book Online Store with Medusa, Gatsby, PayPal, and MeiliSearch

In this tutorial, you’ll learn how to create a comic book store with Medusa and add important ecommerce features.

Tutorial

Medusa Open Source Marketplace Part 2: Make Orders Vendor Specific

In this part, you’ll learn how to link orders to their respective stores in a marketplace.

NEWSLETTER

Receive the most important updates around Medusa and our ecosystem