In my understanding Hyperscan is interesting if you want to check thousands or millions of Regexes against a single string. Normal regular expression libraries would force you to create a state machine for each regex and run them all one by one against a string/byte sequence, whereas Hyperscan builds a single state machine from all the regexes, which it then runs against the string/byte sequence. As most states/regexes will usually be inactive that can dramatically speed up the matching process.

Hyperscan is often used in situations where you have many patterns that you want to test against the same string/byte sequence, e.g. in malware signature detection.

I thought about writing a Hyperscan-alike Regex library in Golang, just not sure how well that would perform with the dynamic memory allocation. I think if you could ensure to do allocations only once during initialization it should be pretty fast. I once built parser generators in Golang and I remember it was a bit painful optimizing them due to dynamic memory allocation, I think for such low-level stuff C or C++ is still better (though I was not very well versed in Golang at the time, so maybe that was on me).