Deployment
A project on your laptop is a demo. A project on a URL is a product.
Most people learning to build stop one step too early. The app runs on their laptop, the demo works, and then it lives forever on `localhost` where no client, recruiter or user will ever see it. Deployment is that missing step, and it is where a portfolio project becomes something you can put a link to. The gap is not talent — it is a handful of concepts nobody teaches in order: what a server actually is, why your API key must never touch your code, how a domain name finds your app, and what to do the first time it returns a 500 to a real person. This module walks that path once, end to end, on your own project. You will push a working AI app to a cloud host, point a domain at it, keep your secrets in environment variables instead of your Git history, and add just enough logging to debug a failure you cannot reproduce. Nothing here is theoretical: by the end you have a live URL you can send to anyone, and you understand every piece well enough to fix it at 2am.
Prerequisites
- You have built a small app that runs on your own machine
- Basic comfort with a terminal and Git
- A project you actually want to put online
What you can do afterwards
- Ship a working app to a cloud host and get back a live URL
- Point a custom domain at your app and understand how DNS resolves it
- Keep API keys and secrets out of your code, in environment variables
- Read a production error you cannot reproduce, using logs
- Explain the difference between your dev environment and production — and why it matters
- Set up a deploy that ships automatically when you push to Git
Sessions
What a server actually is
“The cloud” is somebody else’s computer that stays on. Before touching any dashboard, we build the honest picture: a machine listening on a port, waiting for requests, running your code, sending back a response. Once that is concrete, the hosting choices stop being magic and become trade-offs you can actually reason about.
Covered
Your first deploy
We take your project and push it live. You connect a Git repository to a host, watch the build run, and get a URL back. It will probably fail the first time — a missing dependency, a wrong build command — and that failure is the lesson: reading a build log and fixing it is the skill, not avoiding it.
Covered
Domains and DNS
A random host URL says “student project”. Your own domain says “product”. We buy a domain, point it at your app, and — the part everyone finds mysterious — understand how DNS turns a name into your server. HTTPS comes free once you know where to click, and we cover why a padlock is not optional in 2026.
Covered
Secrets and environments
The fastest way to lose money is to commit your OpenAI key to a public repo — bots scan GitHub for exactly that. We move every secret out of the code and into environment variables, and separate your development setup from production so a test never bills the real account or emails a real user. This is the session that keeps you out of trouble.
Covered
When it breaks in production
It worked on your machine and it is 500-ing for a user in another country. You cannot reproduce it, so you need logs. We add just enough logging and error tracking to see what actually happened, learn to read a stack trace from production, and set one alert so you find out before your users tell you. This is the difference between a demo and something people can rely on.
Covered
Take the project you built earlier and put it fully online: a live URL on your own domain, HTTPS on, every secret in environment variables, an auto-deploy from Git, and one alert wired up. Then break it on purpose — pull a key, trigger an error — and use your logs to find and fix it. You leave with a link you can put on your CV and the confidence that you can keep it running.
Offered in
Questions
Do I need to know DevOps for this?
No. Modern hosts handle the hard parts — servers, scaling, certificates — so a solo builder can ship without a DevOps team. This module teaches the concepts underneath so you understand what the dashboard is doing, not a career in infrastructure.
Will hosting cost me money?
A portfolio-sized app usually fits inside a free tier. The real cost is your AI provider — every model call is billed — which is exactly why we spend a session on rate limits and keeping keys locked down.
I only have a static site — is this still for me?
Yes. We start with static hosting because it is the simplest thing that works, then add a backend the moment your app needs to call an AI model with a secret key — which most do. You will finish able to deploy both.
Other modules
Ready to take this module?