Hackathon Showcase

Xenobiology Station 7

Team led by an AI builder and product designer with 15 years of experience, specializing in AI-first design workflows, gamification, and Chrome extension development.

1 member Watch Demo

YouTube Video

Project Description

Xenobiology Station 7 is a generative creature-building game where Claude is the interface, not a chatbot inside one.
Players pick biological traits; a scientist AI reacts with streaming micro-judgments, builds memory across selections,
and shifts the entire UI’s color atmosphere based on its emotional state. No chat window — Claude’s structured
reasoning streams directly into reactive UI components: trait cards, ambient mood washes, pulsing animations, and a
final specimen dossier with staggered reveal. Stability comes from fallback-first rendering, AbortController-based
stream cancellation, single-component state, and Zod-validated partial JSON parsing.

Team

Drew Miller

Products & Tools

@ai-sdk/anthropic) API routes Anthropic Role: All generation — scenarios Role: Client rendering with useEffect/useState/useRef for streaming state Role: End-to-end type safety from Zod schemas to component props Role: Entire visual system — mood-reactive theming via CSS custom properties Role: Framework Role: Incremental JSON parsing for character-by-character UI updates mid-stream Role: Schema definitions shared between API validation and AI SDK structured output Role: streamObject() / useObject() for structured object streaming from Claude Technologies & Libraries Technology: Claude Sonnet 4.6 Technology: Custom CSS (~800 lines) Technology: Next.js 16 (App Router) Technology: React 19 Technology: TypeScript (strict mode) Technology: Vercel AI SDK v4 (ai Technology: Zod Technology: partial-json ambient data-attribute-driven states gradients micro-judgments server-side streaming endpoints specimen reports trait sets ────────────────────────────────────────

Additional Links

https://station7.vercel.app/

Summarizing URL...