GitHub & Hosting: Where AI code actually lives
AI coding tools (like Cursor or Lovable) just generate text files. To turn those files into a real, live product on the internet, you need version control and hosting. GitHub is the vault that saves and tracks your code history; hosting platforms like Vercel or Supabase are the engines that pull your code from GitHub and run it on the public internet.
A huge hurdle for non-technical founders is realizing that AI doesn't "deploy" itself by default. When an AI writes code, it sits locally on your machine (or in a sandbox). You have to "git commit" (save) and "git push" (upload to GitHub). Once pushed, modern hosting platforms automatically detect the new code and build a live website from it. This entire flow is called CI/CD (Continuous Integration/Continuous Deployment).
Look at the URL of this exact app. It ends in vercel.app. This means the code was pushed to a GitHub repository, and Vercel automatically grabbed it, built it, and hosted it. If you change a file locally, the live site won't change until you push it to GitHub again.
Why doesn't an AI-generated app go live on the internet automatically?