Hey everyone!
I’ve been using WaniKani for a while and kept wishing there was a way to actually read something using the words I’d learned so far. Native content was still too difficult, and most beginner material was too basic.
So I built Mukashi.
You can try a sample story immediately on the homepage — no account and no API key required:
⸻
⸻
What Mukashi does
Mukashi generates short Japanese stories tailored to your vocabulary.
Difficulty modes
- Familiar — uses only vocabulary from your learned WaniKani items
- Easy
- Medium
- Hard
Familiar mode is the one I use most — the idea is that you should be able to read the whole story without needing to look anything up (it’s not perfect yet, but getting better).
You can also:
- choose a topic (or generate one randomly)
- select 3–7 paragraphs
- adjust vocabulary bias toward earlier or later WK levels
If you’re not using WaniKani, you can also generate stories using JLPT vocabulary levels (N5–N1).
⸻
The reader is designed for studying while reading.
- Furigana toggle
- Click any word to see reading, dictionary form, and part of speech
- Kanji tab showing every kanji in the story
- English translation tab (hidden by default)
⸻
Example
If your WK vocabulary includes words like: 公園 ・ 犬 ・ 行くyou might get a short story about going to a park and meeting a dog.
The goal is simply to see familiar words used in new contexts.
⸻
How to use it with WaniKani
- Grab your API key from WaniKani settings
- Paste it on mukashi.app
- Click Save Key
Your API key stays in your browser — it’s never sent to the server.
⸻
A note on AI
Stories are AI-generated and then validated for difficulty and vocabulary constraints.
They won’t be perfect literature, but the goal is exposure. Seeing the words you’re learning used naturally in context.
⸻
I originally built this as a personal study tool, but it reached a point where I thought other WK users might find it useful too.
The app is free and still early access.
Feedback on story quality, the reader experience, or missing features would be hugely appreciated.
Thanks for checking it out,
Brendan
