Fedibook - fighting power with power

As some might know, I'm a strong critic of Big Tech and the giant pile of shit they are bringing upon us. I really, really want "us" to leave the Big Tech platforms and build a better world together — the alternatives are out there.

In 2025 I took it upon myself to step by step find better alternatives and leave Big Tech. I moved my Instagram profile to Pixelfed. I skipped X, never landed on Bluesky, and joined Mastodon. I moved my Scout group from Google, Messenger, and Dropbox to Nextcloud. I told my friends to move to Signal, which many have done. In the end of 2025 I exported my data from Facebook and closed my account.

That last big platform is still the everyday connection point for many, many, many people. It's old — it was once a great tool for keeping in touch, but now it's a giant pile of shit. Also too big to leave.

I love the Fediverse. I love learning new tools. A lot of the great people I connect with in the Fediverse do too — but there is another world of users who are not ready to leave their large network and well-known tools. I tested Friendica. I hustled to bring up Bonfire and also gave Hubzilla a spin. They are great — but I failed to find a setup that I'm ready to propose to everyday users of that F-thing.

After reading the fantastic book Enshittification by Cory Doctorow, it became clear that to fight the network effect and the problem of coordinated move, something else is needed. Something with a very familiar interface and a migration tool.

So here the idea started — what is it we need and can't really find a good, loveable option for? We want a place to connect with friends. We want a way to create and communicate in groups. And we want it to be very easy to use.

I have been building platforms of all kinds since the beginning of the World Wide Web. In 1996 I built the first database-driven intranet and social network (before it was called that) using PHP and MySQL. I have since worked with many available tools — a lot of WordPress — but I'm not a professional programmer. I have worked professionally as a Product Owner with a great team of really skilled programmers who taught me a lot about software design. But building the aforementioned platform as my own hobby project is no way an option.

I have also worked for the last couple of years a lot with AI, using basic prompting — also to write very functional prototypes.

Last week — after having heard so much about Claude Code — a thought flashed through my head. I hate what AI is about to do to the world and how badly we are about to use it — but what if I could turn one bad-ass tech giant against another? What if Claude Code could write the social network to be the killer application in the Fediverse? Let's try!

So I did — and thoroughly impressed I was. Claude Code is so powerful that you can only be both scared and very impressed.

I'm a big believer in fast prototyping — with a fast feedback loop, getting stuff out there and seeing for yourself can work magic, and inspire you to move fast. And boy did Claude Code support this. I started with a conversation in regular Claude about my concept and some requirements. It shall have a familiar look and feel. It shall be possible to bring up on a Docker host with Traefik using a simple .env file and docker compose up. It shall use a tech stack that enables running on a lightweight server and one that Claude Code is good at writing. And it shall support federation using ActivityPub.

Then I went to my dev server, installed the Claude Code CLI tool, and had it build everything — EVERYTHING. The Docker setup, the proxy config, the code, the tools to keep two dev servers in sync. From here there has been a lot of prompting with Claude — but step by step a social media platform took shape, and very early in the process I was so pleased that users joined to test it and give feedback.

After 4 days I was ready to learn how to push the code to the world — again this was very new to me, but I found that Claude had a lot of this covered already. Like maintaining database migration scripts for every change in the database design that had been implemented during the dev process. So now I am able to commit my progress to my new Git repo on Codeberg, for anyone to bring up their own Fedibook instance.

In less than a week I have deployed the showcase instance at https://fedibook.net. Where it will go from here I don't know — I'm super happy to have been able to show my idea and find that it resonates well with the community. I will keep pushing to see where this can go. I'm very well aware that it's a very brute-force process, and that if this actual project is to have a long life, there will come a time where it must move into a more structured approach. That could be a larger developer team, someone taking over the code, or someone just replicating the concept.

For now I will go on. Next up I will have a workshop with Claude where we do some house keeping on the code after this first crazy week.

The first major milestone will be to declare a beta release. A first step toward that will be to establish a roadmap — which for now exists only as some private notes.

For anyone who would like to participate, I encourage you to create an account on https://dev1.fedibook.dk or https://dev2.fedibook.dk to follow the development up close, and to create an account on https://fedibook.net to try out your normal tasks on the showcase server. Of course, with all the risks of running early-stage software — don't use it for anything important.