Inspiration

We wanted to add AI to Live Commerce...

Live commerce has exploded into a $128B industry and the creator economy is expected to reach $480B by 2027. Yet, live auctions and shopping still depend heavily on human hosts, limited schedules, and high production costs. Our team has been big into livestreaming and commerce. Showrunner was inspired by the idea of "AI as the host" using the energy of live entertainment, the trust of authenticated commerce, and the scale of AI automation together into one platform.


What it does

Showrunner is a live streaming auction platform that combines the excitement of traditional auctions with cutting-edge AI technology.
The platform features Luna, an advanced AI host that manages auctions 24/7, providing real-time authentication, market analysis, and engaging bidder interactions.
Built with Next.js and powered by OpenAI, Showrunner transforms how collectibles, art, and unique items are bought and sold online.


How we built it

Frontend

  • Next.js 16 (React 19 + TypeScript)
  • Tailwind CSS for styling
  • Framer Motion for smooth animations
  • Lucide React for icons

Backend & AI

  • Next.js API routes for backend logic
  • Supabase (PostgreSQL) for database, authentication, and real-time updates
  • OpenAI API for AI dialogue and interaction
  • ElevenLabs for text-to-speech voice synthesis
  • Stripe integration for payments

Real-time Streaming

  • WebRTC for ultra-low-latency livestreams
  • Supabase Realtime + WebSocket connections for synchronized chat, bids, and viewer count

Challenges we ran into

  • Synchronizing livestream timers and state across multiple users
  • Handling low-latency streaming in production environments
  • Managing concurrent bid updates and ensuring fairness (anti-sniping)
  • Integrating voice interaction and speech synthesis in real time
  • Achieving stable real-time synchronization between chat, stream, and bidding states

Accomplishments that we're proud of

  • Built a fully functional AI-hosted auction experience in a weekend hackathon
  • Achieved real-time synchronization across all viewers and bidders
  • Integrated AI conversation + text-to-speech seamlessly in a live setting
  • Designed a scalable AI-native livestream architecture deployable on Vercel + Supabase
  • Created an experience that feels like “Twitch meets Sotheby’s — powered by AI.”

What we learned

  • How to manage real-time systems that synchronize state between multiple users
  • The importance of server-time coordination for consistent livestream timers
  • How to integrate OpenAI, ElevenLabs, and Supabase efficiently in one stack
  • How AI can drive engagement and trust in live commerce experiences

What's next for Showrunner

  • Expand to multi-auction rooms and user-hosted sessions
  • Add AI-powered item valuation and fraud detection
  • Enable personalized AI hosts for creators and brands
  • Launch mobile app for live participation and bidding
  • Scale to become the AI-native Twitch for commerce where entertainment, interactivity, and trade converge.

Built With

  • ai-vision-authentication-(planned)
  • context-api
  • edge-caching-(vercel-cdn)
  • elevenlabs-api
  • framer-motion
  • jest-(planned)
  • lucide-react
  • next
  • node.js-20
  • openai-api
  • react-19.2.0
  • react-hooks
  • rls-policies
  • supabase-(postgresql)
  • supabase-auth
  • supabase-realtime
  • supabase-storage
  • tailwind-css-v4
  • typescript
  • vercel-hosting
  • webrtc
Share this project:

Updates