All projects and services that are part of the community ecosystem of starfed.org.
Status: Completed
Media stack for movies, TV shows, and music, powered by Jellyfin, Jellyseerr, Navidrome & Audett (in-house tool).
Status: Completed
A Jupyter server for data analysis, experimentation, and messing around.
Status: Completed
Multiple NAS for file storage and backups for the entire community network.
Status: Completed
Initial containerization of all web services using Docker for easier management and deployment.
Status: Completed
Integrated the Cadaxity and Dyno teams into a single organization, forming the seeds of the Federation.
Status: Completed
Unified login system (SSO) to manage access to all services using LDAP (LLDAP).
Status: Deprecated (archived)
A universal orchestrator and server management tool for all my servers.
Status: Completed
Successfully set up and configured a Matrix server for secure communication, with coturn, and WhatsApp and Discord bridges.
Status: Completed
Transitioned all web services from Docker to Podman for daemonless, rootless container management.
A massive overhaul to rebuild everything from the ground up on new servers with an entirely new architecture. Core services are live, the rest deploys through end of 2026.
Status: Live (core services)
Full infrastructure-as-code with OpenTofu provisioning and Kubernetes (K3s) orchestration, replacing the Podman setup. Every workload is declared in git and synced by a GitOps controller, so the cluster is reproducible from the repository alone.
Status: Live, remainder deploying end of 2026
Services split across several servers in different regions, all federated under one SSO, with scheduled verified backups and a rebuild-from-scratch recovery design.
Status: Live
Account management moved from LLDAP to Keycloak, branded Keystone, giving every service OIDC/OAuth2 login. Identity backups run on a schedule, and restores are verified before the cluster trusts them.
Status: Live
The centerpiece of this epoch. A dead node gets re-provisioned by the automation and re-derives every workload from git and backups, identity included, and the restore is checked before anything relies on it. Rebuild from bare metal is measured in minutes, (its stupid fast), this is still the very early stage so reliability still needs to be mesured but I suspect uptime will be much better than previous epochs, I will setup a status page to mesure uptime after I get Vaultwarden setup. I keep the details off the internet on purpose for security purposes, please ask me in person if you are curious.
Status: Live
Centralized log aggregation with Loki across all cluster nodes and services, shipped by Alloy and retained in object storage.
Status: Planning
Self-hosted password manager for the community.
Status: Planning
Self-hosted email infrastructure for the starfed.org domain.