webOS

• 

• 

Inspiration

I’ve always been fascinated by how operating systems work and how we interact with them. The idea behind WebOS was to recreate that experience—a personal OS in your browser, without needing to install anything, without logins, and with a rich set of AI tools and apps. I wanted to explore whether we could create a fully browser-based, anonymous, app-powered environment that feels like a lightweight desktop OS.

Inspired by platforms like Puter, Windows 11 UX, and emerging cloud workspaces, I aimed to build something lightweight, fast, AI-friendly, and expandable by design.

What it does

WebOS, a futuristic, fully browser-based desktop experience designed to work instantly without installation or login. Imagine a sleek, AI-enhanced OS that works directly in your browser, complete with:

✅ A terminal with custom commands ✅ Built-in LLM assistant ✅ A modern desktop UI (Windows/Linux-style) ✅ Anonymous usage — no tracking or login needed ✅ Web App Store with games, tools, and utilities ✅ Infinite potential for education, developers, and productivity

How we built it

Frontend: Built with React, TypeScript, and TailwindCSS

Window System & Desktop UI: Custom window manager for draggable/resizable apps

Backend: Some parts are serverless, APIs served through Express.js, Socket.IO

AI Integration: Used OpenRouter API to access Claude, Gemini, Mistral, etc.

Streaming: Implemented event-based streaming using text/event-stream to simulate real-time model responses

Proxy Layer: Used CORSProxy.io for bridging models if needed

App System: Apps are stored as modular React components, loaded dynamically

Challenges we ran into

CORS restrictions from OpenRouter and model APIs

Handling streamed responses from AI models correctly in the frontend

Keeping the system lightweight and fast despite dynamic apps

UI layering and focus management in custom windowing

Discovering which model providers were free and stream-enabled

Accomplishments that we're proud of

Created a working OS simulation from scratch

Integrated multiple AI models seamlessly

Designed a modular app system that can scale

Provided anonymous, login-free access with instant usability

Developed a custom socket system for communication and AI tasks

What we learned

How to architect modular web systems that feel like native desktops

How to manage real-time AI interactions in a clean and efficient way

The power of OpenRouter for unified AI access

Limitations of CORS and how proxies can work in web-only apps

UI/UX design considerations for windowed interfaces

What's next for webOS

Add real-time file storage with Firebase or Supabase

Personal login and profile customization

Introduce a marketplace for 3rd-party app uploads

Implement web search with Brave or SerpAPI

Offer collaborative apps: shared whiteboard, multiplayer games

Monetization options like premium AI access, in-app purchases, or team plans

Built With

• axios
• corsproxy.io
• css3
• express.js
• html5
• javascript
• node.js
• react
• socket.io
• tailwindcss
• typescript
• uuid
• vite