Self-Hosted
Self-hosted software put to a decision: measured comparisons, verdicts, and working setups — Docker managers, backups, auth, media, and AI.
Start here
No entry matches those filters. Clear the filters to see the whole list.
What to self-host
12 Best Self-Hosted Apps and Services 2026 to Replace SaaS
The 12 best self-hosted apps to replace SaaS in 2026: Immich, Nextcloud AIO, Vaultwarden, Paperless-ngx, Jellyfin, and more, with Docker files and hardware.
Buying GuideForgejo vs Gitea 2026: Which Self-Hosted Git to Pick
Forgejo hard-forked from Gitea and the two have diverged. How they compare in 2026 on governance, licensing, federation, and migration — and which to self-host.
Comparison · HomelabKarakeep vs Linkwarden: Which Self-Hosted Bookmark Manager (2026)
Karakeep (formerly Hoarder) vs Linkwarden for 2026: AI tagging vs full-page archival, Docker footprint, license, and a clear pick by use case.
Comparison · HomelabSelf-Hosted Discord Alternatives: Can You Host Discord?
Self-hosted Discord alternatives in 2026 compared: Stoat, Matrix/Element, Mattermost, Rocket.Chat, and Zulip. Real RAM needs and the voice reality.
ComparisonWhat to Self-Host First: Six Services in the Right Order
Six self-hosted services in the right order, with real hardware needs: Pi-hole, Uptime Kuma, Jellyfin, Nextcloud, Immich, and Home Assistant.
TutorialMedia, photos and files
AudioBookshelf Docker Setup: Self-Hosted Audiobooks (2026)
Complete AudioBookshelf Docker setup: install, organize your library, set up mobile apps for progress sync, and add podcasts. Replace Audible in under 30 min.
TutorialBest GPU for Plex and Jellyfin Transcoding in 2026
Best GPU or iGPU for Plex and Jellyfin transcoding in 2026: Intel N100 or N150 first, Arc A380 as an add-in card, and a support list for 19 GPU families.
Buying Guide · HomelabBest NAS for Plex & Jellyfin 2026: Which Ones Actually Hardware-Transcode
Which NAS to buy for Plex or Jellyfin in 2026, by transcode load: which chips transcode in hardware, Plex's software rule per chip, dated Newegg prices.
Buying Guide · HomelabImmich Setup and Google Photos Migration Guide (2026)
Move Google Photos to Immich: Docker Compose install, Takeout zips straight into immich-go so dates, GPS and albums survive, then backups and six checks.
TutorialJellyfin Docker Compose Setup & GPU Transcoding (2026)
A Jellyfin 12.4 docker-compose.yml: pinned image, config on SSD, read-only media, Intel Quick Sync or NVIDIA transcoding, Known Proxies and a fix table.
TutorialJellyfin Remote Access: Pick the Right Path for Your Situation (2026)
Jellyfin remote access routed by situation: Tailscale, reverse proxy, VPS relay, or Cloudflare Tunnel - which works on TVs, behind CGNAT, and for family.
TutorialNextcloud AIO Docker Setup Behind Nginx Proxy Manager (2026)
Install Nextcloud All-in-One with Docker Compose behind Nginx Proxy Manager: port 11000 upstream, domain check, Office, Talk, sync clients, Borg backups.
TutorialPaperless-ngx Docker Setup with OCR and Auto-Tagging (2026)
Paperless-ngx 3.5 in Docker Compose with PostgreSQL, Valkey and optional Tika and Gotenberg: OCR scans, tag them by matching rule, back up and upgrade.
TutorialPlex vs Jellyfin (2026): Which Media Server to Run
Plex vs Jellyfin in 2026: what Plex Pass costs now, which one transcodes free, how remote access differs, and how to move a library over.
ComparisonSelf-Hosted Photo Library 2026: Immich vs PhotoPrism vs Ente
Immich vs PhotoPrism vs Ente for a self-hosted Google Photos replacement: encryption, RAM minimums, phone apps, hosted prices and how to leave each one.
ComparisonPasswords, logins and backups
Restic vs Borg vs Kopia: Which Backup Tool (2026)
Choose restic, Borg or Kopia by backup destination. Replay a local restic restore and see why an intact repository can still lack the catalog you need.
Comparison · HomelabSelf-Hosted Password Manager: Vaultwarden on Proxmox (2026)
Install Vaultwarden in a Proxmox LXC with Caddy HTTPS, close signups, hash the admin token, run checked backups and connect the Bitwarden apps. Sync fixes.
TutorialSelf-Hosted SSO Tracker: Which Login Methods Cost Extra?
Compare self-hosted SAML and OIDC by app, edition and price. Check free routes, Rocket.Chat 9 upgrade risks and the sources behind each rule.
Price TrackerDocker, proxies and remote access
Adding a User to the Docker Group Is Not a Convenience Setting (2026)
The command every tutorial gives you to stop typing sudo. What Docker's docs say it actually grants, why re-login is required, and when to use rootless.
TutorialBeszel vs Netdata vs Glances: Homelab Monitoring
Lightweight homelab monitoring compared: Beszel, Netdata, and Glances on footprint, history, and alerts — plus where Uptime Kuma fits, and which to run.
Comparison · HomelabCaddy vs Nginx Proxy Manager vs Traefik in 2026
The three homelab reverse proxies compared: Caddy, Nginx Proxy Manager, and Traefik on HTTPS automation, Docker discovery, licensing, and which to pick.
Comparison · HomelabDocker Cleanup: Which Prune Command Eats Your Data (2026)
Reclaim Docker disk space without losing a database: what system prune removes, why --volumes is narrower than its reputation, and the one to fear.
TutorialDocker Compose Environment Variables: .env vs env_file
The .env file next to compose.yaml puts nothing in your container. See the 6-level precedence table, .env interpolation vs env_file, and config checks.
ComparisonDocker Exit Code 137: Prove OOMKilled Before Raising Memory
Docker exit code 137 is not proof of OOM. Capture inspect state, events, limits, kernel evidence, and cgroup counters before changing the container.
FixDockge Setup: The Compose Manager That Does Less On Purpose (2026)
Install Dockge with Docker Compose, and the one mistake its own README warns about: mismatched stacks paths that write your data somewhere else.
TutorialGlance Dashboard Setup on Windows 11 Without Docker
Run the Glance dashboard as a portable Windows app: verify the download, bind it to 127.0.0.1, start it at sign-in, and fix config and widget errors.
TutorialKomodo vs Portainer vs Dockge: Best Docker Manager (2026)
Komodo for fleets, Portainer for Kubernetes and BE controls, Dockge for one Compose host. Portainer CE stays on 2.45 LTS as 3.0 goes Kubernetes-first.
Comparison · HomelabOllama in Docker with GPU: Stop Installing CUDA Inside WSL2 (2026)
docker run and Compose for Ollama on an NVIDIA GPU under Linux or WSL2, bound to 127.0.0.1:11434 so the open API stays private, checked with ollama ps.
Tutorial · Local AIProxmox vs Docker: Do You Need One, the Other, or Both?
Proxmox runs virtual machines and LXC containers; Docker runs apps in containers. When each alone is enough, and why Docker goes in a VM on Proxmox.
Comparison · HomelabProxmox WireGuard in an LXC: Self-Hosted VPN Setup Guide (2026)
Run your own WireGuard VPN in a Proxmox VE 9 LXC: server setup, phone and laptop configs by QR code, port forward, troubleshooting, vs Tailscale.
Tutorial · NetworkingUptime Kuma Docker Setup, and Where to Actually Run It (2026)
Uptime Kuma 2 in Docker Compose with monitors, alerts and a status page, plus the part guides skip: a monitor inside your network can't see it go down.
TutorialWatchtower Docker: Archived, What It Does, and Alternatives (2026)
Watchtower's repository is archived; the last release was in 2023. What it does, the compose example from its docs, options that matter, and alternatives.
ExplainerLocal AI
Best MCP Servers for a Homelab: 11 Measured (2026)
Eleven MCP servers measured on real hardware: 178-422MB RAM each, ~1GB for a five-server stack. Which earn their memory, which to skip, working configs.
Buying GuideHow to Self-Host a Local LLM on a Single RTX 5060 in 2026
Run Ollama or LM Studio on an RTX 5060 8GB or 5060 Ti 16GB: which models fit, tokens/sec by size, safe LAN access and when to step up.
TutorialLM Studio vs Ollama (2026): Measured Speed and RAM
LM Studio vs Ollama on one RTX 5080: compare the first reply from an unloaded model, 256-token generation, memory use and the settings that changed the result.
ComparisonOllama Alternatives in 2026: Seven Local LLM Runtimes Compared
Seven Ollama alternatives against each project's own docs: licence, platforms, model formats, OpenAI-compatible API, GPU support, and who each suits.
ComparisonOllama vs llama.cpp on an RTX 5080: Measured (2026)
Same GPU, both servers benched: llama.cpp decodes 2-6% faster (more on gpt-oss, via different conversions), Ollama prefills 4-7% faster. One model won't port.
ComparisonSelf-Hosted AI Agent: What It Is and What I Run (2026)
A self-hosted AI agent is the loop, tools and memory on your hardware; the model is a separate choice. The 2026 runtimes, three configurations, and what I run.
ExplainerSelf-Hosted AI and LLMs in 2026: The Stack and How to Start
What self-hosted AI is, the four tool categories (LLM, image, speech, coding), the GPU VRAM each needs, and an Ollama quickstart for an afternoon.
ExplainervLLM vs Ollama: Concurrent Serving vs Local GGUF Swaps
Choose between vLLM PagedAttention continuous batching on Linux or WSL2 and Ollama dynamic GGUF model swapping for 12GB-24GB GPUs and OpenAI APIs.
ComparisonCan people sign in, and can you get the data back?
Decide how users reach the service, what their accounts require and how you will recover their data before switching off the old one.
Before moving a service
- Pick private access or a public endpoint
A private device on your network and a visitor opening a public website need different access paths. Check who needs a client and what is exposed.
- Check the self-hosted edition for sign-in
If users need SSO, check the exact app, edition and release. An included login option in one edition does not establish SSO in another.
- Rehearse recovery before cutting over
In the local restic drill, the images returned but an excluded catalog did not. For your own app, restore its export and metadata to a clean target and try a task a user depends on.
About these self-hosted guides
Self-hosting in 2026 is not a hobby corner anymore — it is the default answer to subscription creep, and the hard part has moved. Installing an app takes one compose file; the real work is deciding which app, what it costs to keep running, and when the self-hosted route genuinely beats the service you are replacing.
That decision layer is what this section is for. Every verdict says who it is for and who it is not, and where a number matters, such as idle RAM, watts or break-even months, I give the number and its method, not an adjective.
Where deployment and access fit
A self-hosting platform is the whole place you run and maintain services. It is bigger than one app.
Start with what to self-host first. Then choose how the services are deployed and updated. For Docker stacks, the Komodo, Portainer and Dockge comparison owns that management decision. For a public web service, the Caddy, Nginx Proxy Manager and Traefik comparison covers the reverse-proxy layer in front of it.
Public access comes next. Use the Tailscale versus Cloudflare Tunnel guide to separate private administration from a site meant for public visitors. For a local AI service, compare model loading, first reply and resident speed separately. One timing cannot answer all three questions.