I want to share an idea for a project I’m developing, and while it’s still in its early stages, I’d love to hear your opinions and suggestions.
The idea is to create a tool that integrates WaniKani and Imabi to help Japanese learners study more efficiently, using two AIs:
-
First AI:
This AI will analyze the learner’s progress in WaniKani (kanji and vocabulary) and Imabi (grammar) to generate a personalized study plan, determining the optimal order of study and combining kanji/vocabulary reviews with grammar lessons.
-
Second AI – Corrections and Adjustments:
This AI will correct and adjust the first AI’s plan (AI can miss to kkkk, it will be for guarantee).
Additionally, I plan to include features such as:
-
Dialogue generation for contextual practice.
-
Integration with Japanese streaming services, allowing students to watch content without subtitles. (I really don’t know if this services will allow that, and of course, you need to have an account)
-
Japanese Manga service like cmoa.jp and BOOKWALKER
-
Customizable settings, where the learner can choose whether they want the AI to focus more on vocabulary or kanji, depending on their needs.
Since the project is still in its early stages, it will likely take a significant amount of time to develop, especially considering I plan to implement a login system and user progress tracking.
I’d really appreciate hearing feedback from the community about:
-
What do you think of the overall idea?
-
Are there any features you’d find particularly useful?
-
Suggestions on how to improve the integration between WaniKani, Imabi, and the additional resources (streaming, manga, dialogues)?
1 Like
Its a cool idea. Sounds like its more a business integration issue than a software issue. So I cant comment on any specifics.
This sounds like an incredible project. I’m not actually familiar with Imabi (I’m currently using the Genki textbooks), but the concept of bridging the gap between kanji and grammar is exactly what’s missing for me right now.
Here are a few thoughts from my perspective:
The “Silo” Problem Right now, I feel like I’m learning in a vacuum. I study kanji on WaniKani and grammar in Genki, but they rarely meet. If your tool could generate practice sentences that use the specific grammar I’ve just learned in my textbook combined with the specific vocab I’ve unlocked on WaniKani, it would be like learning on steroids.
Personalization is everything A common complaint about apps like Duolingo is that the content isn’t relevant to a user’s actual life or interests. I currently have a DIY workflow where I use AI to co-write stories about my hobbies or my last vacation, and then I run them through Narakeet (text-to-speech) so I can listen to them on my commute. If your tool could automate this—creating personalized dialogues that respect my WaniKani level—I’d subscribe in a heartbeat.
Making WaniKani sentences useful I’ll admit I usually ignore the example sentences on WaniKani because they feel disconnected. If those sentences were pulled into the dialogues you’re planning, I’d actually start paying attention to them. I love the idea of using them for listening practice, too—maybe a task where I hear the sentence and have to type the kanji? That would be a great “post-burn” challenge.
One suggestion: Since so many people use Genki or Minna no Nihongo, maybe consider adding an option to sync with those paths alongside Imabi? It would open the tool up to a huge crowd of beginners.
I’m really looking forward to seeing how this develops. It sounds like exactly what I need to finally move past N5!
I want to see an app that lets me use the kanji I learned here and allows me to practice conjugations and grammar. That way I can build on the weak writing skill. Then on top of that integrate something like italki so I can round out and practice speech so I can impress strangers on vr chat. Apparently there are JP-only instances?
Unless you’re very used to produce software for a living I would consider this scope to be way too massive as well as too easy to replace by a simple aggregation of already existing services.
(That said you’re welcome to prove me wrong 
)
If you’re able to produce a single feature that add value to the leaner, put in in production and generate traction it’s another story… but if you look at the similar posts presenting new services using AI on the forum you’ll see that the reception is lukewarm at best.
Best of luck.
2 Likes
My project will be completely free. The AI will use the OpenRouter API, and users will be able to choose which AIs they want to use. The idea is to provide total freedom to personalize the experience, at no cost. If users want, they can use credits on OpenRouter to access better AIs, but it’s not required.
1 Like
About using Genki and Minna, I think I can’t do this because the content is copyrighted. Nihongo might be possible, but I’m not sure.
I can try to create a sync with Genki without the content, just using the titles and user progress. However, for Imabi, with a community-driven document, I can implement a full integration
Italki doesn’t have an official API to integrate. I’m thinking of using OpenRouter Audio Models instead.
I’m trying to do the best possible to make it totally free
About streaming and book services, I’m thinking of just indexing them, but these things are for later. Right now, my focus is on syncing the user data with the server and syncing OpenRouter.
I finished the login and register system with the cryptography now I will sync the login and next think in AI, as I said I’m alone making this so it can take a time because I’m making the interface too
2 Likes
FWIW, I’m interested in watching your progress on this.
I have not yet even begun to dabble with any AI APIs, but I really should start learning about what’s available, including edge AI approaches and tools.
I was not even aware of OpenRouter until I saw your note about using it - and so as a result I will need to jump down that rabbit hole (and others as well).
Your intended ‘bring your own AI to the party’ approach sounds like a great way to go.
1 Like