It might be fast enough though. You haven't specified your corpus size. If your corpus is small enough, ripgrep might be instant enough. Another alternative is to look at fzf paired with ripgrep, but I imagine for large enough corpus sizes, it won't be instant.
Otherwise, if you're searching something large and really absolutely need "real time as I type" as a hard requirement, then you'll not really be looking for a grep per se, but for a tool that builds an inverted index and uses that for searches. So things like Russ Cox's 'codesearch', 'qgrep', 'livegrep' and Sourcegraph.
