Frankenstein AI

Multi-agent AI development platform with horror-themed specialized agents that collaborate to build full-stack applications from idea to deployment.

Comment

Frankenstein AI - Your Horror Dev Team

Inspiration Halloween isn't just about scares—it's about bringing monsters to life! We were inspired by the Kiro Halloween Hackathon to create something that combines the horror aesthetic with cutting-edge AI technology. The idea of assembling a "Frankenstein team" of AI agents, each with the personality of iconic horror villains, felt like the perfect blend of spooky fun and practical innovation. We asked ourselves: What if your development team was literally made up of horror icons? What if Freddy Krueger handled your full-stack development, Pinhead architected your solutions, and Chucky brutally optimized your backend? Thus, Frankenstein AI was born—a multi-agent AI development platform where horror-themed specialized agents collaborate to build production-ready applications. What it does Frankenstein AI is a multi-agent AI development platform that transforms how developers build applications. The platform features a team of horror-themed specialized AI agents, each with unique personalities and expertise:

PinHead (CTO): Solution architect powered by Claude Haiku 4.5 Freddy (Full-Stack Nightmare): Handles complete application development Chucky (Backend Specialist): Brutally efficient with APIs and databases Annabelle (Frontend Specialist): Crafts elegant, creepy React components Jason (DevOps Slasher): Silent but deadly with Docker and CI/CD Pennywise (Testing Clown): Finds bugs hiding in dark corners Ghostface (Security Expert): Paranoid about vulnerabilities

Users interact with these agents through a real-time chat interface to generate development plans, write code, and build complete applications. The agents collaborate, each contributing their specialized knowledge to create structured, thorough solutions for complex development challenges. How we built it Frankenstein AI was built almost entirely using Kiro AI, which became our primary development companion throughout this hackathon. Here's our tech stack and process: Core Technologies

Frontend: Next.js 16.0.3 with React, TypeScript, and Tailwind CSS AI Integration: Anthropic's Claude API (Haiku 4.5) for multi-agent orchestration Animations: Framer Motion for horror-themed 3D effects and transitions Deployment: Vercel with optimized production builds Development Tool: Kiro AI (our main coding assistant)

Development Process with Kiro AI Kiro AI was instrumental in every phase of development:

Architecture Design: Kiro AI helped us design the multi-agent system architecture, suggesting patterns for agent communication and state management. Component Development: We used Kiro AI to rapidly prototype and iterate on 3D horror-themed feature cards with blood-red-to-orange gradients, launch announcement modal with Halloween imagery, interactive agent cards for Freddy, Pinhead, Chucky, and other horror icons, and a real-time chat interface with streaming responses. Styling and Theming: Kiro AI helped us create the signature horror aesthetic with blood-red to crimson to orange gradient theme, supernatural glow effects with dual-layer drop shadows, and dark cards with glowing borders for maximum contrast. Problem Solving: When we hit roadblocks, Kiro AI helped us debug Next.js 16 configuration issues, optimize performance with proper image sizing and lazy loading, fix TypeScript errors and improve type safety, and set up proper Git workflows and deployment pipelines. Deployment: Kiro AI guided us through GitHub repository setup with proper .gitignore for secrets, Vercel deployment configuration, and environment variable management for production.

Challenges we ran into

  1. Next.js 16 Configuration Issues The serverComponentsExternalPackages property was deprecated in Next.js 16, causing our initial Vercel deployment to fail. Kiro AI helped us identify and fix this by updating to serverExternalPackages.
  2. Performance Optimization Balancing beautiful 3D animations with performance was tricky. We initially added GPU acceleration hints and lazy loading, but these caused render errors. We learned to prioritize working code over premature optimization.
  3. Multi-Agent State Management Coordinating conversation history across multiple AI agents required careful state management and message routing logic.
  4. Horror Aesthetic vs. Usability We had to ensure our dark, horror-themed UI remained readable and accessible. The blood-red-to-orange gradient with glow effects struck the perfect balance.
  5. Vercel Deployment Caching GitHub pushed our fix, but Vercel initially cached the old build. We learned to trigger fresh deployments and manage build invalidation properly. Accomplishments that we're proud of

Multi-Agent Coordination: We successfully built a powerful multi-agent orchestration system where specialized AI agents with distinct "personalities" and expertise areas collaborate to deliver more structured, thorough solutions than a single general-purpose AI. Complete End-to-End Development with Kiro AI: We demonstrated the future of software development by using Kiro AI as a true development partner throughout the entire process—from architecture design to deployment—showcasing a real AI pair programming experience. Memorable User Experience: We created a unique horror-themed interface that's not just visually striking but makes AI agents memorable and gives them distinct identities that users can relate to and engage with. Production-Ready Platform: Despite the hackathon timeline, we deployed a fully functional platform on Vercel with streaming responses, real-time chat interface, and a polished UI that balances aesthetics with usability. Innovative Theme Integration: We proved that a horror aesthetic can drive engagement while maintaining professional functionality—our blood-red-to-orange gradient with supernatural glow effects creates an immersive experience that stands out in the AI development tools space.

What's next for Frankenstein AI

Agent Marketplace: Allow users to create and share custom horror-themed agents Code Execution Sandbox: Real-time preview of generated code File System Management: Full project scaffolding and file manipulation Voice Mode: Horror villains that literally speak to you while coding Team Collaboration: Multiple users working with the same agent team Custom Personalities: Users can define their own horror-inspired agent personalities

Built with blood, sweat, and AI Powered by Kiro AI | Deployed on Vercel | Open Source Forever

Built With

  • claude
  • css-transforms-state-management:-react-hooks
  • framer
  • framer-motion-ai-&-backend:-anthropic-claude-api-(haiku-4.5)
  • git
  • github
  • kiro
  • lucide
  • lucide-react-icons
  • monaco
  • next.js
  • next.js-api-routes-ui-libraries:-radix-ui
  • node.js
  • npm-deployment:-vercel
  • radix
  • react
  • react-18
  • react-markdown
  • sandpack
  • syntax-highlighter-development-tools:-kiro-ai
  • tailwind
  • tailwind-css
  • typescript
  • vercel-edge-functions-3d-&-animations:-framer-motion
  • websocket
  • websocket-code-editor:-monaco-editor
Share this project:

Updates