The `.env` file

RUNIQUE_ENV sets how the application behaves; DEBUG only shows the detailed error pages. See Application & Server.

Development (SQLite)

IP_SERVER=127.0.0.1
PORT=3000
RUNIQUE_ENV=dev
DEBUG=true

SECRET_KEY=<generated by runique new>

DB_ENGINE=sqlite
DATABASE_URL="sqlite://data.db?mode=rwc"

?mode=rwc creates the file when it doesn't exist. The URL must start with sqlite://: a short form such as sqlite:data.db is refused ("Unsupported database URL scheme").

Development (PostgreSQL)

IP_SERVER=127.0.0.1
PORT=3000
RUNIQUE_ENV=dev
DEBUG=true

SECRET_KEY=<generated by runique new>

DB_ENGINE=postgres
DATABASE_URL=postgres://runique_user:password@localhost:5432/my_project

Or, instead of DATABASE_URL, the component variables:

DB_ENGINE=postgres
DB_USER=runique_user
DB_PASSWORD=password
DB_HOST=localhost
DB_PORT=5432
DB_NAME=my_project

When DATABASE_URL is set, it wins: the DB_* variables are ignored.

Production (behind a reverse proxy)

IP_SERVER=127.0.0.1
PORT=3000
RUNIQUE_ENV=prod
DEBUG=false

SECRET_KEY=<a key of this deployment's own>

DB_ENGINE=postgres
DATABASE_URL=postgres://runique_user:[email protected]:5432/my_project

ENFORCE_HTTPS=true

The proxy (Nginx, Caddy…) listens on 80/443 and handles TLS; Runique stays on an internal port. For Runique to handle TLS itself with no proxy, see Network & HTTP (ACME).

In production, Runique refuses to start when SECRET_KEY is weak, or when the password reset or the admin is on without .with_public_url(...).

runique new writes one when the project is created. To make another (leak, new deployment): any random string of 32 characters or more.

openssl rand -hex 32

Allowed hosts, CSP, CORS, sessions and middlewares are configured in the builder (main.rs):

.middleware(|m| m.with_allowed_hosts(|h| h.host("example.com")))

See Builder.