What is this?

Kenjanime is a anime fan site developed by individual. You can use a vector search engine to find items similar to your favorite anime and characters.
Vector search is very useful for finding anime with a similar "vibe" or characters with similar characteristics. For example, it's useful for finding more blue-haired characters, or slice-of-life anime with cartoon like character design. It's also useful for looking at lots of anime artwork to refine your design ideas.
Kenjanime generated vectors from anime and character images and created an index aiming to build an easy-to-use recommendation system that allows users to simply click on images they like, without the need for complicated queries.

Vector search has evolved alongside AI advances, becoming highly accurate.
However, the large size of vector data can be quite a problem. For example, if a vector is stored in a database using the 32-bit floating-point format available on many systems, it would take up 3,072 bytes per vector with the currently common 768 dimensions. While this isn't a significant size for general-purpose storage, it still becomes a problem when attempting to "embed" it in a database.
So, with Kenjanime, I decided to quantize it to 1 byte and store it. The vector data size is simply one-fourth the size, at 768 bytes. While this may not seem like a significant difference, it makes a significant difference when you consider that you can store four times as much data within the same budget.
The performance regression doesn't seem to be noticeable, either. At least, it should be sufficient for Kenjanime's purposes.

Yes, half.
Certainly all of Kenjanime's vectors are created using AI. However, because all vectors are pre-computed, they behave deterministically at runtime.
Even so, it can be said that AI is Kenjanime's core technology. Vector search for images that is globally unique regardless of language is undoubtedly an incredible technology made possible by advances in AI.

I'm living in a japanese town where is about 45 minutes by train from Akihabara station of Tokyo. Hello.

It's not particularly popular as a seichi (聖地 anime's place in real), and I don't intend to promote my hometown, but the anime is good, so I recommend you check it out.

I wanted an easy tool to find anime.
If you watch a lot of anime on streaming platforms, the recommendations you get will either be titles you've already seen or new releases. I suppose that's the nature of recommendation systems, but because they're often personalized, they stay the same forever.
This means finding anime becomes a brute force process. I've done it before: I've gone through lists from A to Z, searching for titles that touch me. It's fun at first, but it gets pain, and you end up doing it multiple times because your "mood" changes every week.

The second reason is that I wanted to create something independent of recommendation systems that are linked to reviews and viewing history. Reading reviews is fun, but sometimes you don't want to read them before watching a title. You might also end up reading a review while watching a title and feel sad. Like, "Huh? This one has a low rating? Am I the only one who thought it was good?"
Also, when recommendations are displayed, everyone tends to watch the same titles over and over. That can be fun, of course, but there are times when you want to "dig up" old classics, and I believe that's true for everyone.

In short, I wanted a website that would let users actively use a recommendation system like those found on streaming platforms and shopping sites, only just to find anime. If I had to describe it, it would be a recommendation system without any intention, or a recommendation system where the means have become the end.
And really, Kenjanime is like a streaming platform's recommendation system taken and turned into a website, but it's not personalized, and it doesn't collect browsing history. So, there's no need to log in.

A long time ago, I used to make and post videos on a certain video sharing site. They were comedy-themed. Even though I thought my gags were original, when my senpai (先輩 seniors) would say, "This sounds like that scene from that anime.", I had no idea what they were talking about, and I'd end up like, "Huh? Ah...." and feel sad.
Well, it doesn't really matter, but I think it's sad when we end up saying, "I don't know that anime.". If I could just say, "Ah! I know the name, but I haven't seen it yet. It's an anime similar to XX, right?", I'm sure my senpai would be happy to explain it to me in detail. Well, it doesn't really matter.