Technically, I don't know how to achieve this, but I know that pkgdown and quarto uses some serverless JavaScript solution for this, where the search index is generated when the site/pages are built. I believe they're relying on an established JavaScript package for this, so I doubt we need to roll our own here. I suspect this is a rather simple implementation for someone who already knows how it's done there.
I think the website (https://journal.r-project.org/) would benefit from a search function. It could be free-text search on the content of all https://journal.r-project.org/articles/RJ-yyyy-nnn/ pages. One could later worry about searching also the content of PDFs.
Technically, I don't know how to achieve this, but I know that pkgdown and quarto uses some serverless JavaScript solution for this, where the search index is generated when the site/pages are built. I believe they're relying on an established JavaScript package for this, so I doubt we need to roll our own here. I suspect this is a rather simple implementation for someone who already knows how it's done there.