Stage Side Cafe

Inspiration

When I pictured Twitch meeting the metaverse, I imagined a café or bar where people could gather around tables, watch a stream, and talk with one another. A stream with millions of viewers can make chat hard to follow, leaving people overwhelmed, bored, or exposed to harmful content. StageSide Cafe grew from a simple idea: make livestreams feel more personal, welcoming, and safe.

What it does

StageSide Cafe turns a livestream into a shared social space. Viewers can join small, interest-based café rooms, talk with other fans, and choose their preferred camera angle without affecting anyone else’s view.

For streamers, the platform brings multiple camera feeds into one studio experience. Its AI Sentinel design screens sampled frames in a short rolling buffer and can isolate an unsafe feed while keeping the rest of the broadcast running.

How we built it

We built the web app with React, TypeScript, and Vite, backed by a FastAPI service. Twitch sign-in connects a streamer’s account and channel status to the studio. The system architecture also uses PostgreSQL for account data and is designed to connect Redis, LiveKit, and the media moderation gateway.

The Render demo brings the web app and API together in one service, with a managed database for account data.

Challenges we ran into

The biggest challenge was balancing a fast-moving live broadcast with moderation that can act before unsafe content reaches viewers. We also had to think through how camera switching, chat, and community rooms can work together without disrupting the stream.

Accomplishments that we're proud of

We brought three ideas into one experience: independent camera choice, small-group community, and proactive stream safety. We’re proud that viewers can feel like part of a close-knit room even when the broadcast audience is huge.

What we learned

We learned that live experiences need more than a chat box. Thoughtful group size can make conversation easier to follow, while giving viewers control over their perspective can make a broadcast feel more personal. We also learned how much coordination it takes to design moderation around real-time media.

What's next for Stage Side Cafe

Next, we want to connect the full LiveKit media pipeline, expand real-time AI moderation, and bring café conversations closer to the stream through synced chat and shared moments. We also plan to improve community matching and make the platform ready for larger, persistent audiences.

Built With

• ai-moderation
• computer-vision
• docker
• fastapi
• livekit
• openai
• pgvector
• postgresql
• python
• react
• redis
• render
• twitch
• typescript
• vite
• webrtc