Loading .env with marko-run

Real apps need configuration that changes between environments — a site name, an API URL, a support address, secret keys — without editing the code each time. The usual home for that is a .env file, and marko-run can load one for you. This is a complete, running example — read through it.

The .env file

At the project root:

SITE_NAME=Marko Shop
SUPPORT_EMAIL=help@markoshop.example

Each line is a KEY=value, kept outside your source so it can differ per environment.

Loading it — the -e flag

marko-run does not read .env automatically. The dev script loads it explicitly:

marko-run dev -e .env

-e (short for --env) injects the file’s values into process.env. Leave the flag off and process.env.SITE_NAME is simply undefined. (Under the hood marko-run uses dotenvx; it also picks up a .env.local for values you don’t want to commit.)

Reading it

marko-run renders pages on the server (Node), so process.env is available during SSR. The page reads the config directly:

<h1>${process.env.SITE_NAME}</h1>
<p>Questions? Email ${process.env.SUPPORT_EMAIL}</p>

The preview shows the values from .env, baked into the HTML the server sends — the browser never sees the file.