Backend search with InstantSearch iOS

Frontend versus backend search

Algolia recommends that you use frontend search because:

• It offers a better response speed. Speed alone justifies the choice. Backend search requires two more trips between a user’s browser and Algolia’s servers. This decreases the effectiveness of Algolia’s engine and infrastructure optimizations.
• UI libraries aid development. Algolia’s UI libraries (InstantSearch, Autocomplete, and Recommend) help build a frontend solution so that you can focus on configuring relevance and creating the best possible search experience.
• It frees up your resources. Algolia reduces the need for development and eliminates the requirement for additional server deployments and its associated infrastructure.

Why backend search?

• Server-side pre-processing. Perhaps the most important reason for a backend solution is to make complex calculations or format your data before displaying the results.
• Real-time availability updates (like Airbnb)
• In some situations, SEO is easier to manage.

How to implement backend search

You only need a few lines of code to send a query to Algolia and receive a response. The response contains everything needed to build your UI.

You can perform backend search with either InstantSearch or an Algolia API client.

Backend InstantSearch

Backend InstantSearch offers a good compromise. You perform the search on your servers, carry out any required pre-processing, and then send your results to InstantSearch’s widgets. This enables you to control the query and response but saves you the work of building a custom frontend.

Backend search with an Algolia API client

This option requires a server-side solution and custom frontend development. You manage all queries and responses, and you build the frontend UI. You need to parse the API response: a JSON object. The response includes:

• Results already formatted for display (full record content, embedded HTML, image URLs, and so on).
• Highlighting and snippeting
• Pagination.