I started this out of passion for typing

What an incredibly interesting use of LLMs (generating text to practice typing). It leans in on what LLMs are good at. That said. I would love to see a middle tier pricing which had some features but avoided the AI use.

Hah that's pretty fun. I got tossed about by the animated hands for a few, but grabbed a 194 after that. Dunno about the trigrams though, mostly it's on the "token group" level for me - either the upcoming lookahead feels familiar or it doesn't, and I don't much get bothered by the specific letters as much as "oh I don't have muscle memory on that word, and it's sadly nestled between two easy words, so it's going to be a patchy bit of alternating speed".

This is very neat. One piece of feedback and a gripe I have with a lot of these is that missed or extra characters throw off the entire next sequence and essentially require backing up to deal with them, as opposed to wrong characters which are fine to just be mistakes you move on from. It'd be great to have some detection for when the user is continuing that re-aligns their string.

So I've got some things that seem a little bit weird to me: 1. Typing uppercase characters counts as a mistake I'm not sure how that got to be the case, but somehow typing an uppercase letter instead of the lowercase is a mistake, despite the fact that sentences start with a lowercase letter. This conflicts with my muscle memory of starting sentences with a capital letter. 2. WPM is not a useful metric on its own WPM can rise and fall depending on the length of the word. The bigger the word the less likely you are to type that word correctly from muscle memory, so the speed drops. The speed also drops due to the word being longer. I believe having both metrics would yield more useful data, such as when do you slow down etc. Speaking of which, there are some more statistic things that could help, like measuring how fast you are at fixing the mistakes, or measuring three-letter combination…