[Web App] Yomimaru: A reading app that meets you at your exact WaniKani level, even if you just hit level 5

[klyyze]

This is a third-party app and is not made by the WaniKani team. By using it, you understand that it can stop working or be discontinued at any time.

Hey everyone,

The embarassing part to start. After a couple of years on WaniKani my reviews looked great. Accuracy was solid, my level kept climbing, and the Crabigator seemed pleased with me. Then I’d open a real article, or a LINE message from a Japanese friend, hit a kanji I had genuinely burned, and just freeze. The reading was somewhere in my head. It only ever came out inside a review box, never out in the wild.

That gap is what started this. But the longer I sat with it, the more I realized it wasn’t only a burned-kanji thing, and it wasn’t only a me thing. WaniKani is genuinely brilliant at getting kanji into your head. What it doesn’t hand you is a place to actually read, and finding Japanese to read at your own level is its own headache. Native material is a wall. Most of the stuff written for learners is either dull or buried in furigana for words you already know cold, so I spent the last year building Yomimaru to fix that.

See what it does here

Yomimaru is a web reading app that bends to wherever you happen to be on WaniKani. And I do mean wherever. If you only just hit level 5, this is not a “come back when you’re ready” tool. It will hand you something readable about a topic you actually like, today, at level 5, with the furigana support showing up exactly where you need it and nowhere you don’t. As you climb, the app climbs with you.

How it follows your WaniKani level

You link your WaniKani account once, and from then on every text you open just follows your progress. The kanji you’ve already learned show up bare, the way you’ll meet them out in the wild. The ones you haven’t reached yet keep their furigana, so a sentence is never a brick wall. As you pass and burn more items on Wanikani, or as you keep practicing on Yomimaru, the mastery level of each word is updated, and more characters quietly go bare on their own.

It works just as well early as it does late in your japanese learning journey. At level 5, almost everything still carries its reading, so real Japanese stops being intimidating and becomes something you can actually move through today. The furiganas fade on their own as you earn it.

What you can actually do with it

• Read something at your exact level, today. You can browse a free library of over 10,000 authentic Japanese texts sorted by level and topic, classics from Aozora Bunko and news from NHK, Mainichi, and Asahi. Or, if nothing matches what you feel like reading, have a short story or news piece written at roughly your WaniKani level about a topic you choose (your football team, a game release, a cozy Kyoto mystery). The generation part uses credits, which I’ll explain below. The library never does.

• Read your own stuff too. Paste text or drop in a PDF, EPUB, DOCX, or Markdown file and read it inside the same adaptive interface, with your WK progress applied to it.

• Tap any word, get the meaning and reading inline. No dictionary tab, no leaving the page. Definitions and readings come from JMdict and KANJIDIC, not from an AI guess.

• Translate the sentence that’s tripping you up. When the grammar tangles, pull a full-sentence translation right there, so you’re never stuck staring at a wall you can’t get past.

• Ask the confusing bit out loud. When a grammar point or piece of cultural context trips you up, a built-in assistant (Maru-Sensei) breaks it down in a structured card at your level, instead of a wall of text.

Words you look up save automatically into a review list, and you can export them to Anki if that’s your workflow. Yomimaru is meant to sit next to WaniKani, not replace a minute of it.

What’s free, and where credits come in

The WaniKani sync, the adaptive furigana, the dictionary lookups, the full 10,000-text library, the sentence translations, and Anki export are all on the free Ronin tier. No credit card needed.

Credits only power the creation of brand-new AI content. Reading, looking words up, translating sentences, importing your own files, and running your saved reviews never touch them. And Ronin comes with a generous helping of free credits to start, so you can write your own stories and news pieces at your level without paying anything. If you ever run dry, you simply pause on generating new articles. Everything you were already reading stays open.

What I’d genuinely love your input on

For those of you past level 30 or so: does hiding furigana at Guru feel right, or does an item need to be Burned before its readings should disappear? I don’t have great visibility into what feels natural at every level, and you all do.

And if you’re earlier on, level 5 to 15: does reading with the furigana tuned to your level actually feel doable, or does real Japanese still feel out of reach at that stage? That’s the exact spot I most want to get right.

You can try it here

Thank you for this forum. I’ve quietly read it for years while grinding my own reviews, and a lot of what I learned about studying came from people here. I hope this is a small, useful thing to give back. がんばりましょう

Update (June 8, 2026)

A pretty big round of changes went out since I first posted, so a quick rundown:

• You can now start from absolutely zero. If a friend keeps telling you Japanese looks impossible and they wouldn’t even know where to begin, this is the bit for them. Tell Yomimaru you’re starting from scratch and it walks you through kana first, then hands you your very first all-kana stories, written so you can actually read them on day one. Useful if you ever recommend the app to someone pre-WaniKani.

• A smarter placement quiz. The level check at the start now adapts as you go. It anchors on the kanji you know, then adjusts the vocab and grammar questions to roughly where you’re sitting, instead of throwing random difficulty at you.

• The whole app got a lot faster. Your dashboard, library, discover page and vocab lists now appear instantly on reload instead of showing loading skeletons every time, refreshing quietly in the background. Hard to screenshot, but you feel it right away.

• A guided companion for your daily tasks. Once you start a task from the dashboard, a little floating guide follows you across the app, keeps track of where you are, and lets you mark things done, skip, or jump back.

• News reading now counts toward your stats and reading milestones, same as stories and your own imported texts.

A few smaller things: level-up celebrations no longer pop up in the middle of a chapter (they wait until you leave the reader), the difficulty estimate on a text is more accurate now, checkout supports Apple Pay and Google Pay if you ever try Shogun, and I fixed a handful of mobile reading annoyances. All of the above, including the full beginner kana track, is on the free Ronin tier. Thanks again for following along

Update (June 12, 2026)

Biggest rework Yomimaru has had, much of it straight from feedback in this thread. The short version:

• There’s now a Guided Path, the main way to use the app. Instead of a library and “good luck,” it lays out your next read/review/quiz/roleplay step one at a time, so you never have to decide what to study.

• WaniKani sync moved onto the free Ronin tier (it used to need a subscription).

• The native library is now curated and modernized per JLPT level: archaic spelling cleaned up, genuinely engaging texts hand-scored, tone tags added.

• Reading real Japanese news now counts toward your stats and streaks.

• Plan changes and downgrades live in account settings now, and checkout supports Apple Pay and Google Pay.

Full write-up in my reply below.

A few screenshots of the mobile version

Reading the latest news

Following your learning path (update !)

Keeping track of your progress (update !)

Taking your vocabulary to the next level

Crafting fun content at your Level

Finding actually interesting native content

[neicul]

Just wondering, but are you aware of MaruMori (see thread here: MaruMori.io & Kitsun.io - [Spring Sale ENDING SOON! Mobile App General Release & MaruDex Cross-Platform OCR Lookup Tool Teaser!] Grammar SRS, In-Depth Grammar Lessons, Reading Exercises, Vocabulary, Kanji and more!) which has a similar “teacher” mascot called Maru? Albeit with no AI involved.

[klyyze]

Hey, thanks for the comment!

​I actually used to be a paid Marumori user myself, but since it’s been a while, the mascot’s name (まるちゃん) completely slipped my mind. My apologies for the oversight! I’m definitely going to look into this with my team.

​​On a side note, I’m a huge fan of Marumori, which is why Yomimaru actually tightly integrates with your public API so users can sync their progress between both apps. I genuinely think the two apps work great as a pair. Do you think it would be alright if I introduced Yomimaru in the Marumori Discord server, or would that feel a bit too close to home due to potential feature overlap I’m not aware of ? Completely understand either way!

[DidiBluebell]

I have just tried Yomimaru. First of all, it has integrity issues, for example, if you start a free trial, it automatically chooses the most expensive plan for you. In your profile, there is no way to change this plan either, you have to write to support for that. It feels like scam even if it may be unintentional bad design.

Second problem is that there is no way to click on words in text to see their meaning.

Third problem, it is AI. I got a text on sakura blooming forecast, and it said “By NHK News” without any links to original material. As I chatted with AI, it explained that the weather forecast is fictional and is BASED on real weather forecast. So it is by NHK News really? Sure, for me, as a student, it’sawfully convenient that AI is taking real material and adjusts it to my level. But it’s same stealing as AI stealing from Internet. NHK News generate zero traffic or money from this. Let alone the environmental damage AI entails. I am not against AI, but I am against irresponsible usage of it.

[neicul]

I think you might’ve replied to the wrong person here

[klyyze]

Hi there,

Thanks for taking the time to try out Yomimaru and for sharing your feedback. I want to address your points directly, as a couple of things definitely didn’t hit the mark, while a few others need some clarification.

1. The Subscription Flow

I understand your frustration with how the flow felt, but calling it a “scam” is a bit unfair. The onboarding page explicitly states that the free trial starts on the Shogun plan, and signing up is entirely voluntary. This is a very standard trial model used by the vast majority of software companies, rather than an intentional trick.

That being said, your point about the downgrade process being way too convoluted was 100% valid. Taking your feedback directly to heart, I have just pushed an update to the app : a straightforward, direct way to change or downgrade your plan is now live in your account settings. You no longer have to go through the cancellation steps to switch tiers.

2. Word Lookup Functionality

Regarding not being able to click on words to see their meaning: this is actually a core feature of Yomimaru. If it wasn’t working for you, it sounds like you ran into a technical bug rather than a design choice. Could you let me know what browser or device you were using? I’d love to look into it and get it fixed for you immediately.

3. The “NHK” Text & AI Content

We completely take the L on this one. The text you encountered was one of only two mock content pieces we used during our internal testing phase. They were never meant to be accessible to public users, and leaving them live was an oversight on our part. They have now been completely deleted from the platform.

I want to reassure you that this mock text is absolutely not representative of the app, every other piece of content currently in the global library consists of 100% native literature, sourced directly from the actual Aozora Bunko library (novels, short stories, and essays).

Moving forward, we are strictly focusing on verified, authentic material. Starting next week, we are introducing official news abstracts pulled directly from NHK, NHK News Easy, and Asahi to give users real, properly sourced news to practice with.
We are actively working to iron out these early bumps, and I really appreciate you pointing out where we could improve. I hope you’ll give the actual native library a fair chance!

Best,
Maxime
Founder of Yomimaru

[hrudey]

I took a quick look as I saw the topic and it seems interesting, but the demo example does lead to a question. For reference, I toggled the N3 preset and I’ll provide a couple of screenshots that illustrate the behavior:

I happen to recognize this usage of 中 as a grammar point (shout out to MaruMori) and so since I know the kanji is probably N5 and that grammar point was likely N4 or N3, I was a little surprised to see the furigana, so I decided to look at the details. The details, though, list the なか reading only, and it seems more of a dictionary lookup which is fine (I know Satori Reader by contrast does sentence level lookups and explains grammar points, but that’s a different product also with specifically curated stories and explanations), but not seeing the reading used in the example or even knowing it exists for something that was marked with furigana is surprising. Ideally, all (or at least all common) readings would be listed and the one specific to the selected usage would be highlighted - if the meanings/explanations could somehow be filtered down to that reading or context, that would be even better.

I don’t know if that’s one of the use cases for the AI assistant, but it was an oddity. I read back through the thread and I’m going to admit this is intriguing, but as a developer who has to spend a lot of time disabusing Github Copilot of overenthusiastic “I see the problem, this fixes it” edits that still miss major things that need to be done (and unfortunately am required to use the AI tools) I’d be a little worried about this possibly being an example of the parsing of the documents being done by AI. The reading seems right in this case, though.

In any case, that’s my quick feedback, hopefully it is helpful to you and/or people considering using the application. When I have more time I may take a deeper look, at least to get an idea of the breadth of content available.

[DidiBluebell]

I apologise, must have clicked wrong

[DidiBluebell]

Hi,

Thank you for your answer.

I am using Brave Browser.

[klyyze]

Thanks for taking a look at the demo and sharing such detailed feedback. This is exactly the kind of input we need to improve.

Good catch on the N3 preset showing furigana for 中 on the landing page. To clarify, the landing page demo uses simplified, hardcoded logic. The actual app is much more dynamic: it looks at your WaniKani level, what you’ve read in past sessions, and what you’ve reviewed. In practice, an N3 reader wouldn’t see furigana on 中. That said, it is a weird inconsistency for the demo, and we’ll get it fixed.

On the parsing side, there’s actually no AI involved. We use Kuromoji, a standard Japanese tokenizer, to split the text. For the dictionary, we use JMDict, which is the same database behind Jisho and Yomitan. Right now, the popup is just a direct dictionary lookup. But if you highlight a section and tap Translate or Explain, the app uses AI to give you a context-aware breakdown of the grammar, rather than just raw definitions.

Your point about the popup showing only the なか reading instead of the contextual one is spot on. We are planning to address this by showing definitions based on context. For single-kanji tokens, we want to list all common readings, highlight the one used in the text, and filter the definitions accordingly.

Thanks again for the feedback. It really helps shape the product. Hopefully you can explore it a bit more when things slow down.

[klyyze]

After checking on Brave browser on mobile, clicking on a word to show the dictionary popup works as well as on other browsers. I attach a short video to show how it’s supposed to work
The reason why nothing happened when you tried to click on a word is most likely because you were trying to click on one of the words of the text title or the title of one of the chapters. We apologize for the confusion, and we are going to build the logic to enable dictionary lookup for the titles as well.
Thanks for your feedback,
Maxime.

[gailyn]

How does your product differ from Sartori Reader?

[klyyze]

Great question, and a fair one since there’s definitely some overlap in the spirit of both tools
(reading real Japanese with support instead of just grinding flashcards).

The biggest difference is content. Satori Reader has a fixed catalog. It’s well made, but it’s a finite
set of stories, and for any given level the pool of stuff you can actually read gets pretty narrow. So
you can end up reading a piece about a topic you don’t really care about, or one that feels a little
childish, just because it happens to match your level.

With Yomimaru that problem basically disappears, and it does so on two fronts. First, the library
itself is just much bigger and more varied. We have over 10,000 pieces of authentic native content,
spanning classic literature from the Aozora Bunko archive (Sōseki, Akutagawa, Dazai), real news from
sources like NHK and Asahi, articles, and more, organized by topic and level. So even before you
generate anything, there’s a genuinely wide range of stuff to pick from at your level instead of a
short list.

And here’s the part that really matters: the entire library is free. Not a handful of sample chapters,
the whole thing, no limits, on the free Ronin tier. With Satori only a small slice is free and the rest
sits behind a subscription. So with Yomimaru you get basically endless native reading material without
paying anything.

Then on top of that, there’s content creation. If you can’t find exactly what you’re craving, you just
tell it what you’re into and it generates a story, a lecture, or a news report about that exact thing,
written at your level, in seconds. Love cooking, F1, a specific corner of Japanese history, your
favorite game? You get something to read about it right now.

The news part is the one I’m personally most attached to. You can read about what’s actually happening
in the world today, on any topic you want, written so you can understand it at your current level. So
your reading practice stays tied to real life and stays fresh, instead of being the same closed set of
episodes everyone else is reading.

A few other things that set us apart:

• You can genuinely start from zero. You don’t need to know a single character to begin, because you
  can learn all the kana directly in the app first. Satori assumes you already read kana.
• The reading interface is a lot more customizable. You can shape the look and feel so it gets out of
  your way and you actually sink into what you’re reading.
• For every text you read, you get a comprehension quiz you can learn from, so you can check you really
  understood it instead of just feeling like you did.
• You can practice output too. Maru-Sensei (the built in tutor) will roleplay with you about the story
  you just read, so you get to use the vocabulary and grammar instead of only recognizing it. It’s also
  where you go for grammar and nuance explanations on anything that trips you up.
• Better vocabulary management. Words you look up get saved automatically, and you can organize them
  into different study lists for your reviews instead of one big undifferentiated pile.

Honestly the easiest way to feel the difference is to try it. The free Ronin tier needs no credit card,
and you can dive into the whole library or generate your first piece right away, so you can see if the “read about anything I love” thing clicks for you.

[klyyze]

Hey again everyone, forgot to share this yesterday, but for people willing to try out the more advanced features of Yomimaru we’ve got a promo code for Wanikani users : use the code CRAB during checkout to extend your free trial to one full month !

I’m really curious to get any feedback regarding the AI Content Studio : do the novels, recent news or lectures wrote by Maru-Sensei feel engaging and written in a natural way, or do you prefer to stick to the native content library ?

Thanks for everyone who provided any feedback here, whether it’s positive or negative, it’s really helping us a lot !

[Litehawk]

As a learner I have tried to be extremely selective about how I practice my reading. I try to use only native content because you never know what bad grammar habits you may pick up. Why do you believe that your AI author is capable of writing in an authentic way that is on par with native content? Isn’t there enough native content in existence already? Wouldn’t it be better to focus on curating more tailored existing content than artificially create new content?

[klyyze]

Thanks for this, it’s a really fair question and one I’ve thought about a lot.

You’re right that there’s plenty of native content out there, and I’ll be honest with you: a lot of learners are reluctant to learn through AI-generated content, and I completely understand why. If I could have built this without using AI at all, I would have. But there are a few things worth laying out about why that wasn’t really possible, and why I think the AI concern, while valid, is often aimed at the wrong target.

First, a quick reframe: beginner learning content has always been artificial. Genki, Minna no Nihongo, Tobira, none of that is native content. The dialogues are constructed to showcase grammar points, and sentences like “これはペンです” exist nowhere in real Japanese. Nobody worries those textbooks are corrupting learners, because the real question was never “does this sound like a novel?”, it’s “is this correct?” That’s the standard I think AI content should be held to as well.

When I think “native content,” the first thing that comes to mind is news, and that runs straight into the copyright wall. Buying a license to commercially use NHK or Asahi content would be absurdly expensive. So the only way to give people a free reading experience with genuinely native material is the freely usable Aozora Bunko library, which is exactly what we did. We currently have over 10,000 native texts in our global library, varying lengths, ranging from N5 to N1, all free.

But here’s the catch, and I won’t pretend otherwise: Aozora is very limited for beginners. It has a handful of accessible texts, but the reality is that no matter how much we help with instant dictionary lookups, full-sentence translations, and so on, native content is just genuinely hard for a beginner to read. They need material adapted to their level, and that material barely exists. It’s mostly graded readers (not usable commercially) and NHK Web Easy (very limited, and the content ages out). For someone at N5 or N4 trying to find reading material that’s actually at their level and about something they care about, it’s nothing short of a miracle.

So if we committed to native-only, our options would be:

• Spend a fortune on news licenses and hide everything behind a steep paywall
• Hire people to write content, which means very little of it, and again a steep paywall
• Keep only Aozora, which makes the app useful only to advanced learners

That last one bothers me the most, because it means gatekeeping reading-based learning from the exact people who’d benefit most: beginners.

Now, about the AI reluctance itself, I think there’s one main culprit: the ChatGPT web app. And I couldn’t agree more that it’s unreliable, full of inconsistencies and hallucinations. I wouldn’t dedicate a single hour of my own study time to it either. But there are specific reasons it’s that unreliable, and they don’t apply to what we built:

• The model available in the web app is fairly limited.
• A lot of people tried a model one or two years ago, gave up, and never came back — but the AI landscape has changed enormously since then.
• A more technical point: AI models have a parameter called “temperature” that roughly measures how random or creative the output gets. Most chat apps default it to 1, which is quite high and is the simple reason behind a lot of hallucinations. In Yomimaru, we adjust temperature dynamically per task, for a grammar explanation, for example, we set it to 0.2, which clears out most inconsistencies.
• The most important one: AI will not produce an engaging, immersive story or a natural-sounding news article if you just ask it to write one from scratch. But AI is surprisingly good at copying the tone of an example you give it. That’s the whole gap between the ChatGPT web app and Yomimaru, our novel generation is grounded in actual engaging native novels, and for news, we have a web research layer that pulls the most relevant, recent articles on the topic you want to read about, so the tone of real native articles is matched. We only adapt the difficulty of the language to your level.

That last point is really the heart of it: we’re not asking AI to invent Japanese. We’re asking it to mirror native Japanese and meet you where your level is.

That said, your comment is pushing me toward some native-content angles we’re actively working on or open to:

• News, the legal way: showing only an abstract in-app and redirecting to the original article to read in full. Not perfect, but it sidesteps licensing, and the team is already on it.
• User-imported content: this already works today, you can import a Japanese ebook or paste any text and it’s in your library, fully readable, in seconds. We’re building a browser extension to make this seamless: import any article, blog post, or page you’re reading on the web into Yomimaru with one click.
• Wikipedia: honestly hadn’t thought of this until your comment, but Japanese Wikipedia is free to use. If there’s interest in reading it through our interface, we’ll build it.

So the short version: I’m not claiming AI content beats native content. I’m saying accessible, level-appropriate, engaging native content for beginners essentially doesn’t exist, and our approach grounds AI in real native material to fill that gap, while we keep working to bring more genuine native content into the app.

[klyyze]

Good morning from Japan mates !

Small update on what’s improved in the app since launch:

• Better paged reading mode. We fixed the awkward page breaks and blank spaces that were splitting stories in weird spots.
• The Learning Path got a cleaner, less cluttered look, and it now recommends daily reading tailored to your level.
• More accurate JLPT vocab levels. A lot of words that were wrongly tagged as “unknown” are now classified correctly.
• Sharper word splitting and furigana in the landing page reader demo (the bit you see before signing up).
• It’s now much easier to manage your subscription in the Subscription Settings UI, including downgrading your plan, and unsubscribe from marketing emails directly.

And here’s what we’re working on next:

• Pulling the latest news from Japan’s biggest media (NHK, Asahi, and others) and dropping an excerpt of each article straight into the global library.
• Bringing the most relevant native Japanese Wikipedia articles into the same library.
• A brand new dictionary interface to look up any word instantly inside the app. Plus a more visual overview of your vocab progress, and of your WaniKani/MaruMori account if you sync it.

One last thing, if you have any feedback, ideas, or a feature you’d love to see, please don’t be shy and drop it right here in the thread. We read everything, and a lot of what’s above came straight from your suggestions. You genuinely steer where this goes next.

Thanks for being here, and for all the support so far!

[klyyze]

Hey everyone, it’s been a busy couple of weeks and I’ve got a proper pile of updates this time. Honestly one of the bigger rounds of changes since launch, so grab a coffee.

• The big one: you can now start from absolutely zero.
  Before, if someone showed up not knowing a single hiragana, Yomimaru kind of assumed you could already read a little, which wasn’t really fair to true beginners. That’s fixed now. If you tell us you’re starting from scratch, we take you straight into a guided kana lesson and then hand you your very first all-kana stories, written so you can actually read them on day one. No wall of kanji, no furigana for words you haven’t met yet. So if you’ve got a friend who keeps saying “Japanese looks impossible, I wouldn’t even know where to begin”, this is the thing you can finally send them.

• A smarter placement quiz.
  The level check at the start now adapts to you as you go. It looks at the kanji you know first, then asks vocab and grammar questions that adjust to roughly where you’re sitting, instead of throwing random difficulty at you. And if you ever freeze up on a question, a little nudge reminds you it’s totally fine to skip or come back later. No pressure.

• The app got a lot faster.
  Your dashboard, library, discover page and vocab lists now appear instantly when you reload, instead of making you stare at loading skeletons every single time. It refreshes quietly in the background, so it stays fast and still up to date. This one’s hard to screenshot, but you’ll feel it right away.

• A guided companion for your daily tasks.
  There’s now a little floating guide that walks you through your learning path tasks once you start one from the dashboard. It keeps track of where you are across the app, lets you mark things done, skip, or jump back, and gives you a small celebration when you finish. It makes the “okay, what should I actually do today” question a lot less fuzzy.

• News reading finally counts.
  Time you spend reading the generated news reports now counts toward your stats and reading milestones, same as stories and your own imported texts. There’s also a new daily task to read a real news article if you like having the nudge.

A few smaller things I’m happy about:

Level-up celebrations no longer pop up in the middle of a chapter. They wait until you leave the reader so they don’t break your flow.
Text difficulty estimates got more accurate. The reader is now much better at telling apart real Japanese words that simply aren’t on the JLPT lists from words it genuinely couldn’t read.
Checkout is smoother if you ever decide to try Shogun, with Apple Pay and Google Pay, and it no longer loses your card details if a payment gets declined.
A handful of mobile reading fixes that had been quietly bugging me on my own phone.

As always, all the reading, the whole library, and now the full beginner kana track are there on the free Ronin tier, no card needed. Thank you so much for following along and for all the kind words, it genuinely keeps me going.