Functional diversity is widely used and widespread. However, the main packages used to compute functional diversity indices are not flexible and not adapted to the volume of data used in modern ecological analyses. We here present fundiversity, an R package that eases the computation of classical functional diversity indices. It leverages parallelization and memoization (caching results in memory) to maximize efficiency with data with thousands of columns and rows. We also did a performance comparison with packages that provide analog functions. In addition to being more flexible, fundiversity was always an order of magnitude quicker than alternatives. fundiversity aims to be a lightweight, efficient tool to compute functional diversity indices, which can be used in a variety of contexts. Because it has been designed following clear principles, it is easy to extend. We hope the wider community will adopt it and we welcome all contributions.