Overview

This document focuses on the multimodal search functionality within the search module, which offers a sophisticated search experience by allowing users to search for products using both images and text simultaneously. This feature understands the context of the text query and visual content, providing highly relevant search results that match both the textual descriptions and the visual appearance of products.

Use case

Imagine a user visiting an eCommerce platform to find a "floral pattern dress", but they have a particular style in mind that's hard to describe in words alone. The user can upload an image of a similar dress and add the text query "floral pattern" to specify their interest. The Multimodal Search API processes the combined input and utilizes advanced algorithms to analyze the image's visual cues and the text query's contextual information. It then returns a list of products that closely match the user's described and visual preferences, facilitating a highly accurate and personalized shopping experience.

Example

In the sample request outlined below, we initiated the search for a "floral pattern dress" by an image URL or an image in base64-encoded format, applying faceting based on categories, tags, fit, color, style, and price. We also included specific filters in the search results, ensuring that only items with the fitted fit, and vintage style are included. Additionally, we specified filters to exclude certain results like items tagged as "outlet" and "27/12/2023" or categorized as "jersey" should be excluded. Finally, the sorting is set to arrange the results in descending order based on the item's name.

Sample Request

preparing...

Sample Response

preparing...