NeuralPets

NeuralPets takes the beloved concept of virtual pet companions and reimagines them for the modern AI era, creating a more personalized and customizable experience!

Comment

Inspiration

The inspiration for NeuralPets came when I was thinking back to playing games like NeoPets as a kid. I loved playing various minigames and customizing my pets. These games are are mostly forgotten in today's world, but with the age of AI, I believe these virtual companions can be reimagined. I decided to integrate Gemini-AI to generate an endless number of pets and customization options, and be able to interact and talk with these pets.

What it does

NeuralPets is an AI-powered virtual pet companion. Using Google Gemini AI, it creates intelligent pets with unique personalities that evolve through natural conversations, interactive games (Memory Match, Catch the Treats, Pet Battle), and AI-generated customization. Players build trust and relationships with their pets through chat interactions, earn coins through gameplay, and can customize their pet's appearance using natural language descriptions. The pet's mood, trust level, and responses adapt based on interactions, creating a dynamic, personalized companion experience that grows and changes over time, transforming the simple virtual pet concept into an intelligent, emotionally engaging relationship powered by cutting-edge AI technology.

How we built it

NeuralPets was designed with simplicity and practicality in mind. I built the frontend using React with JavaScript and Vite for fast development, React Router for seamless navigation, and Tailwind CSS for clean, utility-first styling. To keep the architecture lightweight, I avoided a traditional backend and instead used localStorage for pet data and IndexedDB for image storage. For AI capabilities, I integrated Google Gemini 2.0 Flash for conversational AI and personality modeling, plus Gemini 2.5 Flash Image Preview for dynamic image generation, enabling real-time customization and intelligent responses that adapt to each pet's unique mood and trust level.

Challenges we ran into

The biggest challenge was developing the "Catch The Treats" minigame using pure JavaScript and React. Coming from Unity game development experience, this was entirely different as I had to write all the game logic from scratch, from frame-by-frame updates to collision detection systems. Unlike the other games which are more state-based and naturally fit React's component model, this real-time game required manual implementation of game loops, and precise collision detection. The transition from Unity's built-in game engine systems to custom JavaScript game logic required completely rethinking how to handle continuous updates, object movement, and user input in a React environment.

Accomplishments that we're proud of

I'm very proud of having successfully implemented my original scope within the tight timeframe of the hackathon. I was able to deliver on all planned features like AI-powered conversations, interactive games, dynamic customization, and relationship mechanics. In addition this project gave me a way to bring back one of my favorite childhood games and I grew a new admiration for the technical complexity behind the classic virtual-pet experience. Additionally I'm proud of having ventured into previously uncharted territory such as implementing generative AI and real-time game systems in JavaScript, and still be able to overcome these challenge to make a finished product.

What we learned

Through building NeuralPets, I learned a lot about the usage of generative AI API's and also using AI to assist in building the projects themselves. I wouldn't have been able to put out a project of this scope in such a short amount of time if I was writing every line of code by myself, and although it took some trial and error I gained a lot of valuable experience. In addition, I learned how to use JavaScript for more dynamic apps and implementing real-time game logic.

What's next for NeuralPets

First, I'd like to make NeuralPets more responsive so that it can be used on mobile as well. Then, I'd like to expand the scope of the AI within the project, for example giving the chat function context from previous conversations, or having the pet React to your performance in minigames. Talking about minigames, I want to add significantly more games to NeuralPets, as that is the core function of the app, and find other creative ways to integrate AI into it, for example maybe a text-based dungeon crawler with your pet companion. Finally, I'd love to refine and expand the customization options, as that's in my opinion one of the most enjoyable parts of the app!

Built With

Share this project:

Updates