I just found https://www.whatruns.com/ and also found a 3-year-old HN discussion (https://news.ycombinator.com/item?id=15098028) on this tool. I really like it, but I just can't figure out what the operators of that site get out of it? It's free, there's no premium version and they claim to not collect any data. I'm a bit sceptical.
I am actually interested in what goes into building such a product. It would take tremendous resources to crawl millions of websites, not to mention fingerprinting tons of libraries and accounting for all the edge cases...