More and more people are choosing to run their own services at home: private cloud, backups, photo libraries, home automation, media servers, and even local AI models. A homelab today is basically a mini–data center at home, and thanks to open source it’s surprisingly accessible.
Using some of the best projects from the “Awesome Homelab” ecosystem, here’s a practical guide to the tools that make the most sense if you want to start (or level up) your setup in 2025–2026.
1. Foundation: Virtualization, Containers and “Home PaaS”
Almost everything in a modern homelab runs on containers or VMs. A solid base layer makes the rest easy.
Virtualization & storage
- Proxmox VE – A popular open source virtualization platform for running VMs and containers on one or multiple nodes.
👉 https://www.proxmox.com/ - TrueNAS – Perfect if you want a serious NAS: ZFS, snapshots, replication and great support for NFS/SMB/iSCSI to back your containers and VMs.
👉 https://www.truenas.com/ - openmediavault – Debian-based NAS distro, lighter and very friendly for home setups.
👉 https://www.openmediavault.org/
These three cover most homelab scenarios: Proxmox for compute, TrueNAS/openmediavault for storage.
Container management and “click-to-deploy”
- Portainer – A clean web UI for managing Docker and Kubernetes, ideal if you don’t want to live in the CLI.
👉 https://www.portainer.io/ - CapRover – Turns a server into a kind of “Heroku at home”: you deploy apps via a simple web interface or CLI.
👉 https://caprover.com/ - CasaOS – “Personal cloud OS” with an app store feel: point, click, install. Great on mini PCs and small home servers.
👉 https://casaos.io/ - Runtipi – A homelab-friendly launcher with one-command setup and one-click install for common self-hosted apps.
👉 https://github.com/runtipi/runtipi
You don’t need all of them: pick one approach (Proxmox + Docker/Portainer, or TrueNAS SCALE with apps, etc.) and keep it simple at the beginning.
2. AI and LLMs in Your Homelab
If you have a decent GPU (or patience on CPU), you can run local chatbots and agents entirely on your own hardware.
Run models locally
- Ollama – Probably the easiest way to run models like Llama, Gemma or DeepSeek locally with a single command and an OpenAI-style API.
👉 https://ollama.com/ - LocalAI – Open source, local-first alternative to cloud LLM APIs. Acts as an OpenAI-compatible endpoint and can run text, audio, image and more.
👉 https://github.com/mudler/localai
Web UIs for chatting and managing models
- Open WebUI – A modern web interface that supports Ollama and other providers; very popular as a “ChatGPT-style” front-end.
👉 https://github.com/open-webui/open-webui - Lobe Chat – Polished interface for AI agents and chats, with support for multiple providers and knowledge bases.
👉 https://github.com/lobehub/lobe-chat
Agents and visual workflows
- Flowise – Build AI agents and flows visually with nodes and connections.
👉 https://flowiseai.com/ - Dify – Platform for building and deploying agentic workflows, with a focus on production scenarios.
👉 https://dify.ai/ - Cheshire-Cat – Agent microservice focused on extensibility, tools and memory.
👉 https://github.com/cheshire-cat-ai/core
With this combo you can have your own “AI stack” at home: models (Ollama/LocalAI), a web UI (Open WebUI / Lobe Chat) and agents (Flowise/Dify/Cheshire-Cat).
3. Metrics, Dashboards and Analytics
Once your homelab starts to grow, you’ll want to see what’s going on.
Monitoring & observability
- Prometheus – The standard time-series database for metrics.
👉 https://prometheus.io/ - Grafana – The dashboard layer on top: rich charts, alerts, and integrations with lots of data sources.
👉 https://grafana.com/ - Uptime Kuma – A beautiful status/uptime monitor for your services, with notifications when something goes down.
👉 https://github.com/louislam/uptime-kuma - Netdata – Full-stack monitoring with almost zero config; great to quickly see performance issues.
👉 https://www.netdata.cloud/ - Glances – System monitoring in your terminal with an optional web UI.
👉 https://github.com/nicolargo/glances
Privacy-friendly web analytics
If you host websites or apps, you may want analytics without Google:
- Plausible – Lightweight, privacy-focused analytics.
👉 https://plausible.io/ - Umami – Another modern, self-hosted analytics platform.
👉 https://umami.is/ - Matomo – The heavyweight option: very full-featured, suitable even for enterprise cases.
👉 https://matomo.org/ - Metabase – Query and visualize data from your databases without writing SQL.
👉 https://www.metabase.com/ - PostHog – Product analytics, session replay, feature flags and more, all self-hostable.
👉 https://posthog.com/
Together, these tools let you build enterprise-style observability at home.
4. Personal Cloud: Files, Photos and Media
For many people, a homelab starts as a way to escape Google Drive, iCloud or Dropbox.
Files, sync and collaboration
- Nextcloud – The reference for self-hosted “cloud drive”: files, sync clients, calendar, contacts, notes, tasks and a huge app ecosystem.
👉 https://nextcloud.com/ - Seafile – Fast, efficient file synchronization, especially good for big libraries.
👉 https://www.seafile.com/ - Filestash – A web UI for many storage backends (S3, FTP, WebDAV, etc.), nice as a unified file front-end.
👉 https://www.filestash.app/
Photos and videos
- Immich – High-performance, self-hosted photo & video manager with mobile apps and AI-powered search.
👉 https://immich.app/ - PhotoPrism – Feature-rich photo management with face recognition, labels and smart search.
👉 https://photoprism.app/ - LibrePhotos – Another solid option for self-hosted photo management.
👉 https://github.com/LibrePhotos/librephotos
Media centers
- Jellyfin – Completely free and open source media server, similar to Plex but 100 % FOSS.
👉 https://jellyfin.org/ - Navidrome – Self-hosted music streaming server, great as your own Spotify-style solution.
👉 https://www.navidrome.org/ - Audiobookshelf – Perfect if you have a collection of audiobooks and podcasts and want them nicely organized and streamable.
👉 https://www.audiobookshelf.org/
With this stack, you can replace cloud storage, Google Photos and streaming services (at least for your own library) with local, private alternatives.
5. Home Automation and Workflows
If your home has smart bulbs, sensors or switches, the homelab is the perfect place to centralize everything.
Home automation
- Home Assistant – The king of local home automation. Integrates thousands of devices and services, with rich automations and dashboards.
👉 https://www.home-assistant.io/
Visual workflows and integrations
- Node-RED – Low-code flows to wire together devices, APIs and events.
👉 https://nodered.org/ - n8n – Workflow automation tool that connects hundreds of services; like self-hosted Zapier with more power.
👉 https://n8n.io/ - Activepieces – Automation platform built for self-hosting, with strong focus on integrations and AI.
👉 https://www.activepieces.com/ - Huginn – Agents that act on your behalf: watch websites, RSS, APIs and trigger actions.
👉 https://github.com/huginn/huginn
Secure remote access
- Tailscale – Builds a WireGuard-based mesh VPN so you can reach your homelab from anywhere as if you were at home.
👉 https://tailscale.com/ - ZeroTier – Another excellent option for software-defined private networks across devices and locations.
👉 https://www.zerotier.com/
These tools are what turn your homelab into the central brain of your home and your online life.
6. Security and Backups: The Non-Negotiable Layer
Self-hosting is powerful, but it also means you are responsible for security and data protection.
Passwords and secrets
- Bitwarden – Popular password manager that you can self-host or use as a service.
👉 https://bitwarden.com/ - Vaultwarden – A lightweight, Rust-based server compatible with the Bitwarden clients. Perfect for small homelabs.
👉 https://github.com/dani-garcia/vaultwarden - KeeWeb – Client app compatible with KeePass databases, with web and desktop versions.
👉 https://keeweb.info/ - Passbolt – Team-oriented password manager, very useful if you share infrastructure with others.
👉 https://www.passbolt.com/
Backup tools
- Restic – Fast, encrypted, deduplicated backups to many backends (local disk, S3, etc.).
👉 https://restic.net/ - BorgBackup – Another excellent deduplicating backup solution, widely used in the Linux world.
👉 https://www.borgbackup.org/ - Rclone – “rsync for cloud storage”: sync and copy to pretty much any cloud provider.
👉 https://rclone.org/ - Kopia – Cross-platform backup tool with encryption and deduplication, plus a nice UI.
👉 https://kopia.io/
With these you can implement a proper 3-2-1 backup strategy even at home (multiple copies, different media, off-site).
Security and threat detection (for advanced homelabs)
- Wazuh – Open source security platform (XDR/SIEM) for endpoints and cloud workloads.
👉 https://wazuh.com/ - OpenCTI – Cyber threat intelligence platform, great if you want to learn CTI and correlate indicators.
👉 https://www.opencti.io/
These are optional, but very interesting if your homelab is also your security lab.
7. How to Start Without Drowning in Options
With so many great projects, it’s easy to get lost. A sane way to begin:
- Pick a primary goal
- Private cloud and backups → TrueNAS/openmediavault + Nextcloud + Restic/Borg
- Personal AI lab → Ollama/LocalAI + Open WebUI + Flowise/Dify
- Home automation → Home Assistant + Tailscale
- Choose a single base platform
- Proxmox VE for VMs and containers or
- Docker + Portainer on a Linux host or
- TrueNAS SCALE (which combines storage + containers)
- Secure access from day one
- Use a VPN (Tailscale/ZeroTier), not random port forwards.
- Put services behind a reverse proxy like Traefik or Nginx Proxy Manager with HTTPS.
- Document everything
- A small wiki using BookStack or Wiki.js can save you a lot of time when you forget how you configured something.
👉 https://www.bookstackapp.com/ – BookStack
👉 https://js.wiki/ – Wiki.js
- A small wiki using BookStack or Wiki.js can save you a lot of time when you forget how you configured something.
If you tell me what kind of homelab you want to prioritize (media, AI, backups, smart home, “all in one”…), I can sketch a concrete mini-architecture with 4–6 apps from this list and how they fit together on one or two machines.