Modern society faces a growing loneliness epidemic. To address this, we developed an interactive dating simulator that lets users input basic personal information and receive tailored character matches—using familiar Clash Royale personalities. Users can then chat with these characters through a dynamic, AI-driven interface. Our goal is to provide comfort, entertainment, and a playful social experience through an accessible digital platform. On the technical side, our simulator is a full-stack web application. The frontend is built with VSCODE, where users submit a lightweight profile. The backend, implemented in Node.js / Python / other, offers REST and WebSocket endpoints to handle profile submission, matching, and rich chat.
We designed a matching engine that maps user traits to a vector space of Clash Royale characters; each character has a defined personality profile, and we compute similarity scores to pick the best match.
For the chat system, we use a hybrid dialogue simulation: characters’ responses are generated via a pre-defined tree structure, but smarter responses come from a lightweight language model that's prompt-tuned for each character’s voice.
We store user profiles, chat logs, and character dialogues in a MongoDB / PostgreSQL database, making it easy to maintain conversation context and refine character personalities over time.
The system supports secure communication over HTTPS, with persistent user sessions and data encryption.
Log in or sign up for Devpost to join the conversation.