Developers from ZLVOX have introduced AI Anime Finder — a service that allows for searching anime based on natural language and "vibe" descriptions. The tool leverages neural network capabilities to match complex emotional queries with the AniList database via GraphQL.
What Happened
The AI Anime Finder project allows users to enter text descriptions of moods or plots—for example, "cyberpunk romance with enemies to lovers"—and receive relevant results from AniList. In addition to the search engine, the service includes a watch-time calculator that accounts for fillers and a generator for visual "Taste Cards" for social media sharing.
Context
Traditional search systems in media databases usually rely on rigid categories and tags (genres, studios, years). However, such filters are often unable to convey abstract concepts or specific emotional states that are important to users when selecting content.
Why It Matters for the Industry
The project serves as an example of the successful implementation of LLMs and semantic search in niche user interfaces. It demonstrates the potential of using a "semantic wrapper" pattern around existing GraphQL/REST APIs to create deeply personalized experiences in specialized domains such as music, books, or video games.
Why It Matters for Users
For viewers, this means a transition from searching by dry genres like "fantasy" or "action" to searching by specific feelings and scenarios. Users gain the ability to find content that precisely matches their current emotional context, which is difficult to achieve through standard filtering systems.
What Is Not Yet Known / Limitations
At the moment, the project appears to be a high-quality demo; there is no data regarding the system's scalability, model inference costs, or the overall production-readiness of the solution.
Sources
Author
Look at AI, Editorial Team