InstantSearch from Algolia is incredibly fast search
We started using Algolia this past spring, and until the last 6 weeks or so, made relatively little use of the expansive features. In early November, we started making much more use of the platform's capabilities. It's allowed us to reconfigure the entire workings of a project and increase the speed immensely.
For one projects, Algolia is now used to store records for over 35,000 active products and delivers results in a matter of seconds. The same index is available in a number of different interfaces, all pulling nearly instantly from a central database.
Search results are highly configurable, both in the display and the ranking. It's interesting to watch the results respond nearly instantly to updates to the way results are scored. Instantearch.js gives a straightforward option to allow users to filter searches by configuring faceted fields in the backend.
New beta features give the options to both suggest synonyms by 'AI' and automatically re-rank items using an AI filter. In our products database, there's a lot of variation in the way products are described and tagged (i.e. women's shoes from a focused shoe brand vs women's shoes from a clothing brand that also sells shoes) and it will be interesting to see if this can help with some of the issues. We're also working on implementing a machine learning algorithm to help with better tagging and will have more info on how that impacts the search results soon.
Algolia has been a fantastic platform to build a database of many thousands of records, and serve them in a lightning fast manner in a number of environments. The customer support has been incredible as well, and we're excited to make more use of Algolia's newest features!