You gotta appreciate the commitment to getting a usable thing in place: being able to actually just... replay the requests--it's an impressive thing.

I wonder if the project's purism impairs its ability to be repurposed, though? One of the cool things about web archival projects is how through content-hashing / IPFS type nonsense, we might be able to build systems that let us have a shared memory of the internet. This seems important for images/media where duplicating that content is a heavy lift. (I'm thinking a bit of Jortage here, which makes it so a single-user Mastodon instance does not incur separate storage of every image it comes across...) It's hard for me to picture this project fitting into a shared archive -- but then, that's only my angle on web archival, and I haven't picked through the tech here.