This part of the documentation is exclusive to starters that use meilisearch

Enterprise Commerce leverages Meilisearch’s hybrid search technology to provide a powerful and flexible Vector Search feature. This allows users to go beyond keyword-based searches and find relevant results based on semantic meaning and conceptual understanding.

​
How it works

Vector Search utilizes AI-powered embeddings to represent data points (e.g., product descriptions, customer reviews) as vectors in a high-dimensional space. Similar data points have vectors that are close together, enabling the search engine to identify and retrieve conceptually relevant results even for complex or nuanced queries.

​
Benefits

  • Improved Search Relevancy: Vector Search significantly improves search result quality by understanding the meaning behind queries, not just the specific keywords.
  • Customizable: Users can fine-tune the search experience by adjusting the “semantic ratio” to prioritize keyword-based or semantic search depending on their needs.
  • Future-proof: The underlying technology is constantly evolving and improving, ensuring that Enterprise Commerce remains at the forefront of search capabilities.

​
Setup

Enabling Meilisearch’s Hybrid Search in your Enterprise Commerce project is a straightforward process:

  1. Navigate to your Meilisearch Cloud dashboard.
  2. Select your project.
  3. Enable the “AI-powered search” feature.
  1. Go to the “Indexes” tab and access your default index settings.
  2. Open the “Embedders” tab.
  1. Click “Create new embedder”.
  2. Name your embedder “default”.
  3. Provide your OpenAI API key.
  4. Set the “Document template”. This template determines what information from each document is sent to the AI model for embedding generation.
A product titled '{{doc.title}}' worth '{{doc.minPrice}}' , whose description starts with {{doc.description|truncatewords: 20}}

Note: You can customize the document template to fit your specific needs and data structure.