I saw your blog post describing the algorithm used in this app and I wonder about this difficulty factor. In Anki there seemed to be the whole issue around "ease factor", so much that one of the most used addons modified this. It's an outdated issue with the new FSRS there though.

So what I'd like to ask is, how does your current algorithm behave when you fail a review many times, then start being succesful again? Do the initial failures mean you're forever cursed with short review interval for that specific review, or will it fix itself? I see

I've finished all the WK kanji in Ringotan and I've found that hitting Ringotan on "frequent review" accelerates my learning so much that I really want to do it with every new non-Jinmeiyou kanji I find in the wild. My first question is, is one of the filter options when selecting what list you want to study kanji from comprehensive, i.e. includes everything that's inside of Ringotan? Because I keep finding myself having to jump from list to list to find different kanji I come across. (a) Is one comprehensive; (b) if not could we get that :); (c) if

There is tons of great open source data out already, like https://github.com/chanind/hanzi-writer?tab=readme-ov-file and https://github.com/skishore/makemeahanzi and given you already have Ringotan put together, I wonder if feeding this data in wouldn’t be fairly trivial. I can ALMOST replicate Ringotan-like function using this tool https://github.com/krmanik/Anki-xiehanzi to pull from those resources into an Anki deck, but by default they're teaching hanzi by having you draw over the outlines instead of testing you on hanzi like Ringotan does with kanji. I spent a few hours trying to duct tape my way into a solitution, to mixed results. But even if I managed to get