I made a (completely free!) tool called sakubun – it helps you practice kanji, and improves your vocabulary. You can check it out here: [https://sakubun.xyz](https://sakubun.xyz)
First, it lets you create a list of kanji that you’ve learnt so far. You can do this by importing an anki deck, or from your WaniKani account, or you can also add it from a predefined list like RTK or the JLPT (or any webpage, really – you can copy paste). You can, of course, also add kanji individually.
Sakubun can then give you quiz questions where you have to figure out the kana readings of sentences that use _only_ the kanji that are in your kanji list. This is useful to learn new vocabulary that use the kanji you already know, and you’ll get more used to reading those kanji from increased exposure.
It also has an essay feature – you can generate “essays”, which are collections of random sentences that use only the kanji from your list, and you can practice reading them together. These essays can be read either horizontally (like English) or vertically (the way Japanese is written traditionally).
Additionally, it has a custom text feature – you can type/paste Japanese text from anywhere, and it’ll add furigana to all the words that have kanji you haven’t learnt yet. This way, you can practice reading text from the real world, but use assistance only for kanji that you haven’t learnt so far.
The quiz has a collection of over 130K sentences, taken from the tatoeba project. Despite the original tanaka corpus having more sentences, any sentences that were marked unreliable are not included here.
The website works on all devices, and can also be installed as an app if you’d like – instructions are provided in the website. It’s completely free and open source software, and the backend is written in rust. Please feel free to contribute if you’d like! You can find the GitHub repo [here](https://github.com/cubetastic33/sakubun).
Any queries, suggestions, or feedback of any kind is _highly appreciated_!
1 comment
Based on the errors that popped up in the 5 or so sentences I tried I assume you’re using MeCab with IPAdic and no (or very little) post-processing for generating the readings in the custom text tab. You should switch to UniDic, since that generally has much better results even on its own. IPAdic probably will contain one error per sentence or two, and more if it contains rare kanji