What's already running out of the box
A one-click deploy brings up all four required containers: the Immich application server, the machine learning service handling face detection and CLIP search, a PostgreSQL instance with the vector extension it depends on for similarity search, and Redis for caching and job queues.
Uploaded photos and videos are written to a persistent volume separate from the database, so the media library and the metadata database can, in principle, be backed up independently. On first startup, the machine learning container downloads its inference models before facial recognition and smart search become fully functional — this can take a few minutes and is normal, not a broken deploy.
First login and admin setup
Immich doesn't ship with a preset admin account. Instead, opening the server's web address for the first time lands you on a setup wizard where you create the admin account yourself, entering an email and password of your choosing — there's no default credential to look up in a panel. The web interface defaults to Immich's standard port unless Leneex's template maps it differently.
Connecting the mobile app
After the admin account exists, install the Immich app on iOS or Android and point it at your server's address to start automatic backup — the app version and server version should stay reasonably close together, since large version gaps between them are a common source of sync errors after updates.