Chaos Challenge — Your Stream. Their Choices. AI-Powered Chaos.

Inspiration

Twitch is all about connecting streamers with their communities, but audience participation is often limited to chatting, voting, or using channel points. We wanted to make livestreaming more interactive by allowing viewers to influence what happens during a stream. That's how Chaos Challenge was born — an AI-powered Twitch application that turns viewer suggestions into fun, unpredictable challenges for streamers.

What It Does

Chaos Challenge allows Twitch viewers to participate directly in a stream by suggesting and voting on AI-generated challenges. Key features:

• AI-Generated Challenges: Creates challenges based on viewer suggestions and stream settings.
• Live Voting: Viewers vote through Twitch chat to decide the next challenge.
• Chaos Memory: Remembers previous results and adjusts future challenges.
• Streamer Dashboard: Allows streamers to manage and approve challenges.
• OBS Overlay: Displays challenges, voting results, and countdown timers live on stream.

How We Built It

We used OpenAI Codex to help develop the application through a step-by-step development process, from planning and coding to debugging and testing. The application uses:

• React, TypeScript, and Tailwind CSS for the frontend.
• Python and FastAPI for the backend.
• Twitch API and EventSub for chat integration.
• Ollama for locally generated AI challenges.
• SQLite for storing challenge history.
• WebSockets and OBS for real-time updates and stream overlays. We focused on using free and open-source tools to keep the project accessible.

Challenges We Faced

The main challenges involved integrating real-time Twitch chat, managing audience votes, and keeping the dashboard and OBS overlay synchronized. Another challenge was ensuring that AI-generated challenges were creative, safe, and relevant to the stream while keeping the application lightweight.

What We Learned

This project helped us understand how AI can make livestreaming more interactive and engaging. We also learned more about Twitch API integration, real-time communication, local AI models, and how OpenAI Codex can simplify the software development process.

What's Next?

In the future, we hope to introduce multiplayer streamer challenges, AI voice announcements, channel point integration, and automatic gameplay detection. Our goal is to make Chaos Challenge a fun and interactive experience where streamers and their communities create memorable moments together.

Your Stream. Their Choices. AI-Powered Chaos.

Built With

• fastapi
• github
• obs-studio
• ollama
• openai-codex
• python
• react
• sqlite
• tailwind-css
• twitch
• twitch-eventsub
• twitch-oauth
• typescript
• vite
• websockets