Write your first Caddyfile

Shortcuts are fine for experiments. For anything you want to keep, you write a Caddyfile, a small text file that describes your sites.

The Caddyfile exists because Caddy’s native configuration format is JSON, and nobody wants to write that by hand. The Caddyfile is a human-friendly layer that Caddy converts to JSON for you.

Create a file named Caddyfile in an empty directory:

:8080 {
  respond "hello from a Caddyfile"
}

The first line is the site address. It tells Caddy what to listen for. :8080 means: any hostname, port 8080, plain HTTP. Everything inside the braces defines how that site responds.

Run it in the foreground:

caddy run --config Caddyfile

Then test it from another terminal:

curl http://127.0.0.1:8080/

While you’re learning, always run Caddy in the foreground with caddy run. You see startup errors and log lines the moment they happen, and Ctrl-C stops the process cleanly instead of leaving a mystery server running in the background.

Why did :8080 give us plain HTTP? Because Caddy has no name to issue a certificate for. Change the site address to a hostname like blog.example.com and Caddy switches to automatic HTTPS for that site. That behavior gets its own module later in this course.

Now the mistake everyone makes on day one: you edit the Caddyfile and nothing changes. A running Caddy does not watch the file. You have to stop and start it, or better, tell the running server to pick up the change with caddy reload. We’ll build a safe reload habit in the operations module — for now, remember that saving the file does nothing by itself.