Hello again everyone!
I’ve been a user for a long time, although I took a fairly large break when my daughter was born a few years ago before restarting my Japanese journey fairly recently.
I wanted to share a small side project I’ve been building for a while (I started it during my paternity leave a couple of years back but I haven’t had time to finish it until now) to help me bridge the gap towards immersion. I struggle a bit with immersion particularly because between work and my daughter I don’t have a lot of time to sink into watching a lot of movies/shows.
I’ve built two parts to help me with that:
-
A video player built for immersion, with grammar analysis, embedded dictionary and sentence+audio+image extraction with an SRS system (and export to Anki)
-
A related website that generates small comprehensible input stories from the key vocabulary to be able to get started with a particular piece of content you want to watch. For example let’s say you want to be able to watch Ponyo. Ponyo has a total vocabulary of 740 words, but to get a comprehension of around 70% you only need 190 of those. At 40 words a week that’s 5 weeks to be able to start watching Ponyo on your own and then naturally acquire the rest of the vocab passively.
Both systems are connected, so you can review the SRS from the web or the app and the SRS can mix both sentences from the generated stories as well as from the content itself.
The story generation is LLM-based, but I’m putting a lot of effort to constrain the LLM to generate good, entertaining, natural sounding and comprehensible stories at the uer level with the appropriate grammar and register to understand your target content properly. I don’t claim to have nailed this yet but I think I’m getting there and most of my current effort is focused there.
One thing I want to add soon is audio for the generated stories using TTS (I know I know, it’s not the best, but creating personalized dynamic content without it is impossible). I think since the idea is to get you rolling with actual real content ASAP (instead of replacing that) it’s not a completely bad solution.
In the ideal case scenario I help you bootstrap your comprehension and you stop using my app altogether! The story generation and SRS is paid but the video player will be fully free. I also want to add an export to Anki so you can always get it out if you don’t like it or want to stop paying.
The immersion video player is not fully ready yet (I think it should be in 1-2 weeks), but the website can be used now at https://katarineko.com/. I’ve created a discount voucher for 50% off forever for the monthly or annual plans for WaniKani users with the WANIKANI promo code and there is a 14-day free trial.
I would love any feedback, especially:
- Does this sound like a useful bridge into native content?
- Does a film-specific vocabulary target feel motivating?
- Would you prefer importing known words from WaniKani, or using a separate calibration?
- Does the overall idea make sense, or am I overcomplicating the problem?
If anyone wants to try the story generation in exchange for honest feedback, I am happy to offer a few free Pro subscriptions.
Also let me know if something on the page isn’t clear, my copywriting is the worst.
