I think you need to start self hosting
Sigh.
Long time no see.
Anyway, let’s get into it.
I think you need to start self-hosting. And no, before you scroll away, this isn’t another post telling you to spend thousands of dollars on servers or become a Linux wizard.
Hear me out.
First, what the hell is self-hosting?
You know all those applications you use every day? Google Photos, Google Drive, Notion, password managers, cloud storage, automation platforms, media servers, and about a million other subscription services?
Most of them are Software as a Service, or SaaS.
Essentially, someone else runs the software on their infrastructure, maintains it, and gives you access, usually for free with limitations or in exchange for a recurring subscription.
And that’s perfectly fine.
But what if I told you that you could run alternatives to many of those applications yourself?
That’s self-hosting.
Instead of depending entirely on someone else’s infrastructure to run your applications and store your data, you take responsibility for running the software.
And the interesting part?
You don’t even need an actual server to get started.
That old laptop sitting in your wardrobe collecting dust? It could be your server.
A mini PC? Absolutely.
A virtual machine on your existing computer? Also works.
You could even rent a cheap VPS from a cloud provider and host your applications there.
The point isn’t necessarily owning physical hardware. It’s having more control over the software you depend on.
And what can you self-host?
Well, a ridiculous amount of things.
Want your own Google Photos alternative? Look at Immich.
Your own cloud storage and file-sharing platform? Nextcloud.
A personal Netflix-like media library? Jellyfin.
Your own password manager? Vaultwarden, an unofficial Bitwarden-compatible server.
Want to automate workflows without paying for every little operation? You can self-host n8n.
Need uptime monitoring? Uptime Kuma.
Want to block ads and tracking domains across your home network? Pi-hole or AdGuard Home.
You can host development environments, Git repositories, dashboards, websites, databases, VPN services, home automation, and even your own AI tools.
The rabbit hole goes very, very deep.
And before anyone starts shouting about AI, yes, you can self-host language models too.
Tools like Ollama make it possible to run compatible open-weight models on your own hardware.
Your conversations can stay on your machine, you can experiment without paying per API request, and you get more control over how your models are deployed.
But let’s be realistic.
Frontier models from companies like OpenAI, Anthropic and Google generally remain more capable across many demanding tasks than the smaller models most people can reasonably run at home.
And running large models locally requires serious computing resources, especially if you want good performance.
So no, I’m not suggesting you abandon ChatGPT and somehow recreate a frontier AI laboratory in your bedroom.
Use the right tool for the job.
But here’s where things get interesting.
AI has made self-hosting more accessible than it used to be.
A few years ago, if you wanted to deploy an unfamiliar application, configure a reverse proxy, troubleshoot containers, understand networking, and automate backups, you’d probably spend hours reading documentation and Stack Overflow threads.
You still need to understand those things.
But today, AI coding assistants, reusable automation scripts, and tools like Docker Compose can make the process significantly easier.
You can ask an AI assistant to explain configuration files, help troubleshoot deployment errors, generate scripts, or walk you through setting up a service.
You can even build small applications tailored specifically to your needs.
Something that would normally require subscribing to yet another SaaS product might be relatively simple to build and run yourself.
And that’s what excites me.
Not simply saving money, but having the ability to decide how the tools I use should work.
Because here’s something people don’t always consider.
We’ve gradually become tenants in our own digital lives.
Your files are on somebody else’s platform.
Your photos are on somebody else’s servers.
Your notes, workflows, projects, and sometimes even your passwords are managed through services whose pricing, policies, features, and availability you don’t control.
A company can increase subscription prices.
It can discontinue a feature you rely on.
It can impose new limitations.
It can shut down entirely.
And suddenly you’re scrambling to find an alternative or export years of your own information.
Now, self-hosting doesn’t magically eliminate these risks.
But it can reduce your dependence on a single provider.
It gives you greater control over your applications, your data, and how everything connects.
And for someone like me who enjoys tinkering with infrastructure, that’s an incredibly satisfying feeling.
Of course, there’s a catch.
You’re now responsible for keeping the damn thing running.
Congratulations, you’re the IT department.
Something breaks at 2 AM? That’s your problem.
A security vulnerability gets discovered? You need to patch it.
Your hard drive dies and you didn’t configure backups?
Well, I hope those files weren’t important.
Self-hosting means thinking about security, backups, updates, storage, networking, availability, and disaster recovery.
And if you’re exposing services to the internet, please don’t just open random ports and assume everything will be fine.
Use proper authentication, HTTPS, secure remote access, and sensible network restrictions.
Also, let’s kill another misconception.
Self-hosting isn’t always cheaper.
Electricity costs money. Hardware costs money. Internet connectivity costs money. Cloud servers cost money.
And your time isn’t free either.
Sometimes paying $5 or $10 monthly for a managed service is objectively the better decision.
There’s absolutely nothing wrong with that.
This isn’t about turning everything into a DIY project just because you can.
It’s about having the option.
Personally, I think the best approach is hybrid.
Self-host the applications where privacy, customisation, ownership, learning, or cost make it worthwhile.
Use managed services where convenience, reliability, and advanced capabilities matter more.
You don’t have to replace everything overnight.
Start small.
Install Docker on an old computer.
Deploy Uptime Kuma or a simple dashboard.
Learn what containers are, how persistent storage works, and how to access applications securely over your network.
Then maybe try Nextcloud, Immich, or n8n.
Experiment. Break things. Fix them. Learn.
And please, set up backups before trusting your new setup with important data.
The more you experiment, the more you’ll realise how much of the software you use every day can be run on infrastructure you control.
And you might eventually find yourself building an entire personal ecosystem of services that work exactly how you want.
One application becomes three.
Three become ten.
Before you know it, you have a homelab, several Docker Compose files, a monitoring dashboard, an unhealthy obsession with networking, and a server you refuse to turn off.
Welcome to the club.
But jokes aside, I think there’s something genuinely valuable about understanding the technology we depend on.
You don’t need to become a systems engineer.
You don’t need to abandon every commercial product.
And you certainly don’t need to build your own data centre.
Just understand that renting access to software isn’t always your only option.
Sometimes you can run it yourself.
Sometimes you can modify it.
Sometimes you can build something better suited to your needs.
And in a world where practically every application wants another monthly subscription, I think that’s a skill worth exploring.
Own what makes sense. Rent what makes sense. But at least know you have a choice.
Anyway, that’s enough preaching for today.
Would you guys be interested in a series where I show you how to build a self-hosted setup from scratch, using affordable hardware and free software?
Because trust me, there’s a lot we can do with an old computer.