OTAKULT is an anime recommendation engine that learns from every choice you make. No search bars. No trending carousels. No algorithm designed to keep you comfortable. Just the show you need to watch next. Whether you know it yet or not.
You swipe through anime. It watches. Every like, every skip, every obsessive rewatch pattern you reveal. The engine builds a vector of exactly who you are as a viewer and uses that to surface what's next.
Genre tags are a starting point, not the answer. OTAKULT goes further. It knows the difference between an action show you want for spectacle and one you want for emotional gut-punches. It knows that two shows can share the same genre and deliver completely different experiences. It pays attention to that.
The longer you use it, the more accurate it gets. That's not a feature. That's the whole point.
It started as a personal problem. Anime catalogues are enormous, recommendation algorithms are garbage, and the people who actually know what to watch are spread across Reddit threads, Discord servers, and obscure forum posts from 2009.
OTAKULT started as a 2AM project by me, built on strong opinions about recommendations and enough stubbornness to do something about it. I hope it becomes something people actually use, so I can justify making it even better.
We process real community review data. Not star ratings, but what people actually write about what a show made them feel, what it cost them in sleep, what it changed about how they watch anime. That language becomes signal. Signal becomes your taste profile. Your taste profile becomes the recommendation you actually wanted.
Every anime in the catalogue has a signal vector, a set of values derived from how thousands of real viewers described the experience: emotional intensity, binge pull, visual craft, darkness, pacing, divisiveness, comfort factor, and more.
When you swipe, those signals combine into your personal taste profile. The engine continuously re-ranks the entire catalogue against that profile, using cosine similarity across weighted dimensions. It's not magic. It's math applied to real human language about what anime actually feels like to watch.
The directory lets you browse all 18,000+ titles by genre, tag, and popularity if you'd rather explore on your own. Every page links back to the engine so you can take what you find and let it seed what comes next.