Deploy a Flask app built with AI

Deploy a Flask app, step by step

1. Create your account at project3000.io with Google or email, and set up two-factor sign-in (it takes a minute).
2. Choose a package and create the app. Its name becomes its address: your-name.apps.project3000.io.
3. Open your Flask project folder in an AI coding tool (Claude Code, Codex, Cursor or Gemini CLI), with approval turned on, so it asks before it runs each command. The “See how” links in the dashboard show where both settings are.
4. Copy the prompt from your dashboard and paste it into your AI tool. It reads your project, adds the two files Project3000 needs (a Dockerfile and a manifest.json), connects the app to its included database and cache, and uploads it. Approve each step when your AI tool asks.
5. Click Deploy. Your app goes live on its own HTTPS link and runs as two copies on separate servers. Share the link with anyone.

What changes for Flask online

• Gunicorn (or a similar server) runs the app instead of flask run.
• The server must listen on 0.0.0.0 and the port in manifest.json, not localhost: Flask’s app.run() listens on 127.0.0.1 unless told otherwise.
• SQLite becomes the included PostgreSQL-compatible database, through DATABASE_URL (SQLAlchemy works).
• Sessions are kept in the cache or signed cookies, so both copies know who is signed in.

You don’t make these changes by hand: the deploy prompt explains them to your AI, which checks the project with the same validator our servers use before it uploads.

Packages

Prices in US dollars, before VAT or sales tax. Billed monthly per app; cancel anytime from your dashboard.