AI is changing the world. Don't stay behind. Clear summaries, community insight, delivered without the noise. Subscribe to never miss a beat.

What I Self-Host

viceconsole.fyi

October 15, 2026

16 min read

🔥🔥🔥🔥🔥

70/100

Summary

A self-hosted home server built around a Beelink S12 Pro mini PC runs 28 services on an Intel N100 processor, 16 GB of RAM, a 512 GB SSD and a 6 TB media drive. The machine, bought for €180 last year, uses about 7.5 GB of memory and 30% to 40% CPU according to htop. It does not run local AI inference. Services are accessible only on the local network or through VPN; VPN endpoints run on the router and are publicly routable only over IPv6. Jellyfin replaced commercial streaming subscriptions and supplies locally controlled movies, television and home videos without autoplay or recommendation feeds. The setup also includes Immich for photo storage, Home Assistant for automation, Frigate with eight IP cameras, Vaultwarden for passwords, Actual Budget, Forgejo, FreshRSS, Audiobookshelf, Syncthing, Samba and Uptime Kuma. Frigate uses lower-resolution camera streams to limit Wi-Fi, processing and storage demands, while Jellyfin generally uses 720p media to reduce resource use. Seventeen services run natively and 11 run in Podman containers, all managed through systemd. Native deployments are preferred when dependencies are limited, while Podman is used for more complex applications. The server avoids Docker, the *arr media-management stack, self-hosted email, Collabora Online and several database-heavy alternatives because of operational complexity or resource demands.

Key Takeaways

A Beelink S12 Pro with an Intel N100 CPU, 16 GB of RAM and 6 TB of attached media storage runs 28 self-hosted services while using about 7.5 GB of memory.
Jellyfin replaced paid streaming services and provides locally hosted media without autoplay or recommendation feeds.
Services are not exposed directly to the public internet; remote access relies on VPN servers on the router that are reachable over IPv6.
The deployment uses 17 native services and 11 Podman services managed by systemd, with Docker no longer used.
Immich, Frigate, Home Assistant, Vaultwarden, Syncthing and Uptime Kuma provide self-hosted photo management, camera monitoring, home automation, password syncing, file replication and availability monitoring.

What the discussion said

The AI-adjacent part of the thread centered on whether self-hosted products should accept heavier infrastructure when it enables genuinely useful intelligent features. Immich became the clearest example: one reader objected to every service demanding its own PostgreSQL and Redis stack, arguing that self-hosted software should stay lean enough to run on SQLite. Others pushed back that Immich’s database choice is not gratuitous complexity; its semantic photo search depends on VectorChord, letting people search a library by the content of an image rather than filenames and folders. For them, that is exactly the kind of capability worth a modest resource bill. Commenters also treated AI-assisted coding as a practical catalyst for personal software. One participant said it had led them to run more homegrown applications than packaged ones, while another mentioned a personal finance tool built this way. The enthusiasm was tempered by irritation at the surrounding ecosystem: a complaint about today’s flood of low-effort AI-generated software framed the fragmented media-server stack as an embarrassing failure to deliver an integrated experience. Still, no one disputed that locally run AI features and AI-assisted development are making self-hosting more capable and more personal.

Where opinion split

The sharpest disagreement was whether Immich’s PostgreSQL-based stack is unjustified overhead for self-hosters or a necessary price for semantic image search. Critics want simple, portable deployments built around SQLite; defenders argue that searching photos by concepts such as objects and scenes requires vector-search infrastructure, and that capability is worth the extra services.

Read original article

Community Sentiment

Mixed

Positives

Immich’s vector-backed photo search turns a private library into something people can query by visual meaning, making self-hosted AI feel useful rather than ornamental.
AI-assisted coding is lowering the barrier to bespoke personal tools, with commenters already running more custom applications instead of waiting for polished products.

Concerns

The infrastructure needed for semantic search frustrates self-hosters when each application insists on isolated PostgreSQL and Redis services, turning a useful AI feature into deployment sprawl.
The complaint about an era of low-effort AI-generated software carries over to self-hosting: users still lack a cohesive media tool despite abundant code and automation.