We highly recommend front-end search.
A front-end implementation hits our servers directly, without going through your back end. It can be up to 10x faster for the end user to retrieve results.
We provide a distributed API, allowing you to select multiple data centers around the world. This lets your users target the closest server, which has a significant impact on latency. This can only work with a front-end implementation.
With a back-end implementation, you’re forcing some network routes. Even if there’s only a local problem in the network, this affects all users. Using a front-end implementation means that only a subset of your users are affected when there’s a network issue.
What is front-end search?
What type of user experience you create is entirely flexible—Algolia can power a wide array of potential search patterns. When you query Algolia, the engine returns a JSON response, which allows you to consume the search results easily and craft any experience. The response contains a breadth of useful information beyond matching results. You get access to the total number of matches, pagination, computed facets, highlighting, and more.
Take a look at our InstantSearch UI libraries for more detail, including code samples and a widget showcase.
So, while your server remains responsible for many Algolia-related activities—like importing and updating data or configuring indices—it doesn’t need to get involved in the search experience.
If you feel like a back-end implementation would still work better in your case, don’t hesitate to get in touch with us.
Article is closed for comments.