For larger data sets, generating index and strigifying it as JSON from a backend service is faster to load into the FE. Check [lunr docs](https://lunrjs.com/guides/index_prebuilding.html).