Yeah the Threadreader is really just for threads, it doesn't capture any replies. Nitter and Ghostarchive do that.

So an archive of a thread unrolled by TRA looks like this: https://archive.is/znrBS

A Ghostarchive of a thread looks like this: https://ghostarchive.org/archive/SYuCV - captures replies as well

And a Ghostarchive of a single tweet looks like this: https://ghostarchive.org/archive/TtGaQ - also with replies

Note that Nitter instances nor TRA are not actually archives, they will error out if the source tweet disappears, you still need to archive them somewhere. It's particularly stupid with TRA, it would be easier to just scrape the thread and save it locally instead of loading it from twitter every time since they don't care about replies, but maybe it has something to do with copyright law, I have no idea.

Also, Ghostarchive has the big advantage of not driving me fucking insane with captchas. Archive.today gives me a shitty fucking captcha nearly every time I try to either archive something or view an archive, and I'm getting to the point where I want to run whoever owns it with a bus, beat him to death with a bicycle, and impale him on a fucking fire hydrant next to a flight of stairs. Jesus fucking Christ. That's why I occasionally shill it, I wish people would switch from .today.