fast-autocomplete | Demo app that shows how to implement a fast autocomplete | Autocomplete library

Hello I have an app that have autocomplete suggestions for the user based on input in the search bar.

I use a package named fast_autocomplete which works great but I have a problem that each time I want to use the prediction I have to load 50MB of JSON data that I mad for the autocomplete (500K records). While running it on a server side loading, parsing and sending back the data is quite slow for what you would except for autocomplete functionality. It less than a second and right now (testing locally) its takes few seconds.

Checking the issue it seems that it takes a lot of time to load the 50MB JSON for each request. Loading the data and building the autocomplete object each time a new request comes in is a waste of time.

For that I wondered if there is a way to keep the object loaded alive all the time and when a new HTTP request comes in the JSON is already loaded. How big sites like Amazon, Ebay, Google makes there autocomplete so fast?