Inspiration
We were inspired by the growing demand for user-generated content and the gap between professional game studios and everyday creators. While social platforms empower people to post videos and art instantly, building interactive games still requires deep technical expertise. We wanted to remove these barriers and make game creation as simple as writing a prompt, opening up opportunities for hobbyists, educators, and brands to engage through play.
What it does
Volibear is an AI-powered platform that lets users create, share, and play games instantly. By entering a natural language prompt, users can generate playable web-based games powered by large language models. These games can be shared via a social feed, discovered by others, and even customized for brand engagement. The platform democratizes game creation, fosters community-driven experiences, and creates new opportunities for interactive marketing.
How we built it
We combined multiple layers of technology:
- LLM Game Generator: Uses an open-source Claw LLM fine-tuned for structured game logic.
- Supabase Backend: Manages user data, game storage, and feed interactions.
- React Native App: Provides a cross-platform interface with mobile and web compatibility.
- Web Game Hosting: Ensures scalable deployment and instant playability.
- Social Layer: Built around a dynamic feed where users can publish, remix, and engage with games.
Challenges we ran into
- Designing a reliable prompt-to-game pipeline that outputs structured, playable game code.
- Integrating Supabase for real-time social feed updates while maintaining performance.
- Handling cross-platform compatibility for both mobile and web.
- Balancing AI creativity with constraints to ensure generated games are playable and not broken.
Accomplishments that we're proud of
- Successfully built an end-to-end pipeline from text prompts to live playable games.
- Deployed a working cross-platform prototype accessible via web and mobile.
- Integrated a social feed that enables sharing, discovery, and engagement around games.
- Established early frameworks for brand integration, showing the commercial potential.
What we learned
We learned how to leverage LLMs for structured content generation beyond text, the importance of designing feedback loops for AI outputs, and the technical challenges of building cross-platform apps with dynamic, user-generated content. We also deepened our understanding of scalability and real-time database architectures.
What's next for Volibear
Our roadmap includes:
- Launching a public beta with creator tools and SDKs for developers.
- Expanding the game template library to cover diverse genres and mechanics.
- Building stronger brand integration APIs for interactive marketing campaigns.
- Scaling to multi-platform support with larger communities and monetization opportunities.
Built With
- agentic-llm
- ai
- babylon.js
- cloudflare
- docker
- finetune
- gcp
- groq
- langchain
- mcp
- mongodb
- next.js
- nginx
- node.js
- openai-oss
- phaser.js
- react
- react-native
- supabase
- tailwind
- three.js
Log in or sign up for Devpost to join the conversation.