Yeah, so far I’ve mostly used koohii. You can export the words for a given book. Koohii has a limited selection for books, so recently I created a webscraper to get the words from http://jpdb.io, which is a similar site. But that doesn’t have everything I would like to read either, so I spent some time fiddling with Japanese language tokenizer to make a (very rudimentary) tool to extract a word .csv from a text.
You can export the list of trashed words from koohii so I just filter those out. I’ve used floflo/koohii for enough time that I’ve trashed a lot of words, so it works pretty well. I still find myself suspending about 50% of the words that end up in Anki, though. Also, Anki doesn’t import words of a similar note type that already exist.
You don’t have to prelearn the whole book. You could just learn 20 words and then read to the point where that takes you, since the words are chronologically ordered. What I did in the beginning was to only prelearn the frequency 3 words, since learning all the frequency 1 words would take forever, and those can be quite rare and difficult words. I can look those up as I go if I feel like it. I still don’t prelearn freq 1 words with all books; I’m reading All You Need is Kill ATM and since sci-fi is a genre I haven’t really touched in Japanese, there’s like 1k new unique freq1 words .
Anki is a flashcard application available for desktop (Windows, macOS, Linux). It is primarily offline and syncing to online is possible, but optional.
There are also mobile versions for iOS (AnkiMobile, I think) and Android (AnkiDroid). iOS version is also made by the Anki creator, but Android is made by a different team.
Personally, I use Anki on Windows and Linux and AnkiDroid on a smartphone and a tablet.
Admittedly, I tried to create Rep2recall, which uses not only a fully online database (no syncing problems), but also utilizes Handlebars / EJS templating language, which probably more powerful than mustache-ish templating used in Anki (where I can’t totally replicate as well). A goal of the project was to allow programmatic manipulation as well.
@neicul Any thoughts? A part of my idea is about programmability, which may interfere with security. Another thing is note taking and handwriting (with digital pen). I am also thinking about individual card override.
Thanks, everyone! I’ve taken ideas from everyone who replied (so I couldn’t really mark any one answer as the solution, sorry), and I’m starting to configure my deck(s) in a way that is more to my liking. (It’s quite the learning curve, but it seems to be worth it). I now wonder if there’s an easy way to add native audio to my cards (and if so, whether it’s even a good idea).
I 100% recommend adding native audio if at all possible. At least for me, this is the only way I have any hope of my flash cards translating into listening comprehension (though I’m someone who struggles with auditory processing, so it might be more optional for you!).
I have two main methods of doing it: the first and easiest method is letting Yomichan add audio for me, since the majority of cards that I create through Yomichan will automatically import an audio file of the word when a card is created from it.
The second method is what I use to add audio to a deck I downloaded from someone else, or occasional cards that Yomichan does not have an audio file for. I use the Forvo pronunciation downloader Anki add-on, which makes it pretty painless to add audio to a card. I’ve come across words that neither Yomichan nor Forvo has audio for, but so far it has only been a small handful.
Thanks, I didn’t know I can import sound directly from Yomichan! I followed this guide from sycamore’s post above and it doesn’t import sound. I’ll have to see how to go about adding it (still very much learning). I’ll also look into this add-on.
Yomichan actually adds audio by default (at least I don’t remember having to change it? The link to the audio file shows up in the reading field for me), but admittedly I’ve only barely modified the contents of my cards (I configured it to pull the sentence the word was found in, but that’s it). I bet you could find somewhere on the card to put the audio (or create a new field for it) and then mess with Yomichan’s settings so that it knows which field to send it to. However, it might make your cards uglier, since the audio usually comes with a play button. I’ve styled mine a little with CSS so it’s more elegant looking, haha
Wow, I just checked out that Refold website. I like it! I’m going to try some of their philosophy/approach/recommendations to start reading some Japanese for enjoyment (and ‘tolerating the ambiguity’) rather than ‘perfect understanding’ (trying to look up every word or grammar point I don’t understand). I previously tried following a beginner’s book club on the Sailor Moon mangas, but even that got me way too frustrated, since I was hyper-examining every word and phrase trying to completely understand every sentence – when what I could have just done instead is to read to the best of my (current) ability and just enjoy the gist of the story for the nostalgia and fun of it!
I’m also following an approximation of their roadmap, but I’m close reading a few pages of よつばと！ every day and looking up a bunch of grammar and words as well. For enjoyment I’m watching 日常 without worrying too much about comprehension, and just to acclimatize to the sounds of Japanese I’m listening to the podcast あらB.fm
Yay, I added an audio field and now audio is automatically loaded to the card. Plus, if I re-add an existing word it seems to fill in the missing audio? (not sure if there’s an easier way to update all of them, but they’re relatively few so far, so it’s no big deal). The Forvo add-on also works like a charm. I wouldn’t know how to tweak the play button’s appearance, but I don’t mind it as it is. Thanks again!
.
I’ve taken ideas from everyone who replied (so I couldn’t really mark any one answer as the solution, sorry), and I’m starting to configure my deck(s) in a way that is more to my liking. (It’s quite the learning curve, but it seems to be worth it). I now wonder if there’s an easy way to add native audio to my cards (and if so, whether it’s even a good idea). 
(I configured it to pull the sentence the word was found in, but that’s it). I bet you could find somewhere on the card to put the audio (or create a new field for it) and then mess with Yomichan’s settings so that it knows which field to send it to. However, it might make your cards uglier, since the audio usually comes with a play button. I’ve styled mine a little with CSS so it’s more elegant looking, haha