Estuary TypeScript SDK Demo

Real-time AI voice and text conversation demo built with the Estuary TypeScript SDK (@estuary-ai/sdk) and Next.js 16.

What This Demos

• Text chat with streaming responses
• Voice input via WebSocket or LiveKit WebRTC
• Live speech-to-text transcription
• Animated voice orb with state-aware visual feedback
• Bot interrupt capability
• Connection lifecycle management

Getting Started

npm install
npm run dev

Open http://localhost:3000 and enter your Estuary credentials:

• Server URL: https://api.estuary-ai.com (or your self-hosted instance)
• API Key: Your est_... key from the Estuary dashboard
• Character ID: UUID of the AI character to chat with
• Player ID: Any unique identifier for the end user

Tech Stack

• Next.js 16 (App Router, Turbopack)
• React 19
• Tailwind CSS v4
• @estuary-ai/sdk
• livekit-client (optional, for WebRTC voice)

Project Structure

src/
  app/
    page.tsx          # Entry point (dynamic import, SSR disabled)
    layout.tsx        # Root layout with fonts
    globals.css       # Theme, animations, Tailwind config
  components/
    ChatApp.tsx       # Main chat interface (config screen + chat UI)
    VoiceOrb.tsx      # Animated orb visualization
  hooks/
    useEstuary.ts     # React hook wrapping EstuaryClient

Dogfood Findings

See DOGFOOD_FINDINGS.md for documented painpoints, bugs, and wishlist items discovered during development.