Requirements
The machine
Section titled “The machine”| Resource | Base setup |
|---|---|
| Memory | About 350 MB in use on a fresh instance (API 210 MB, web 85 MB, PostgreSQL 40 MB). Plan 1 GB. |
| Disk | About 2 GB for the images, plus the database, which grows with what people track, and backups. |
| CPU | One core is enough; work peaks during imports and the catalogue refreshes. |
| Docker | Docker Engine with Compose v2 (docker compose, not docker-compose). |
Each optional service adds its own containers on top; monitoring, with Grafana, Loki and Prometheus, is by far the heaviest.
Architecture
Section titled “Architecture”The images are built for x86-64 (amd64) and 64-bit ARM (arm64);
Docker pulls the right one on its own. That covers VPSs and ARM cloud
servers, mini PCs, NAS models, Apple Silicon Macs, and a Raspberry Pi 4 or 5
running a 64-bit system.
32-bit ARM (armv7, older Raspberry Pis or a 32-bit system) isn’t
supported.
Systems
Section titled “Systems”Anything that runs Docker Compose works the same way: a Linux server, a NAS with a Docker app, or Docker Desktop on Windows and macOS for a try. The installation needs a shell on it, to clone the repository and run two commands.
On a NAS, keep the repository on a volume that is backed up, since .env
and your keys live there, and point the addresses at the NAS: see
Installation.
Network
Section titled “Network”- Outgoing: the server reaches the catalogues (TMDB, AniList, IGDB, Open Library, MusicBrainz) live on every search, and the GitHub Container Registry for updates.
- Incoming: ports 8080 and 3000 on a home network, or 80 and 443 with a domain and HTTPS.
- A domain with HTTPS is required to install the app on a phone and for push notifications.