Stop Renting, Start Owning. Build a Social Media site. It's the New Blog.
Building in public, part 3. Build your own, I'll show you how. If you're old enough, you remember having a blog. Blogger and LiveJournal both launched in 1999, and WordPress followed in May 2003. You picked a theme, wro
Building in public, part 3. Build your own, I'll show you how.
If you're old enough, you remember having a blog. Blogger and LiveJournal both launched in 1999, and WordPress followed in May 2003. You picked a theme, wrote whatever you wanted, and people found you through RSS and blogrolls. It was small, but it was yours.
Then most of us made a trade without really noticing. We moved our writing, our photos and our arguments onto platforms that had the audience. The deal was simple: they host it for free, and we get reach.
This piece is about why that deal has gone bad, why AI has made the old idea of owning your space practical again, and how to actually do it. I'll use Chirpper, the social site I've been building since March, as the worked example, including the parts that haven't gone well.
The deal we took, and how it went
The reach part of the deal got worse every year.
Facebook. In 2012, a Page post reached about 16% of the Page's followers on average. By March 2014 that was down to about 6.5%. Recent benchmarks put it under 2%. A page with 10,000 followers now reaches roughly 200 of them for free.
Google Reader. Google shut it down on July 1, 2013, citing declining use. RSS never fully recovered as a habit, and a lot of independent blogs lost their main way of being read.
Reddit. In 2023 Reddit started charging for API access that had been free since 2008. Popular third-party apps like Apollo shut down, and more than 8,000 subreddits went dark or read-only in protest. The moderators who built those communities for free had no say in it.
The pattern is always the same. You build the audience, the platform owns the room, and the platform changes the rules when it needs revenue. None of that is evil. It's just what renting looks like.
What changed: building one is now cheap
For most of the last 15 years, "build your own social network" meant raising money and hiring a team. That's no longer true.
I sent my first prompt for Chirpper to ChatGPT on March 5, 2026: a social site with no login or signup that still works as a great social site. Two days later, on March 7, a working version was live in production: you could post, generate invites, and reply. I'm one developer working with AI tools on nights and weekends.
The hosting is cheap too. Chirpper runs on Cloudflare Pages with a D1 (SQLite) database. Cloudflare's paid Workers plan starts at $5 a month. The free tier gives you 5 million database row reads and 100,000 row writes a day, but since September 1, 2026, queries on the free plan fail once you pass those limits, so budget the $5 once real people show up. [CONFIRM: Chirpper's actual monthly hosting bill]
The code is not the hard part anymore. That changes the question from "can I build this?" to "should I?"
Why a social site and not just a blog
A blog is a broadcast. You write, people read, and maybe a few of them leave comments that nobody else reads. The conversation about your post happens somewhere else, on someone else's platform.
A social site puts the conversation in your room. Your readers can post, reply to each other, and bring their friends. Think of it as an alternative to a reddit subreddit, that you own: the community, the moderation rules, the archive and the domain are all yours.
It also gives your own writing a home with structure. Every Chirpper member gets an articles page with covers, tags and an RSS feed. That's a blog, just inside a place where other people also show up.
Should you build one? An honest checklist
It's a good fit if:
You already have a group somewhere you don't control. A subreddit you moderate, a Facebook group, a Discord, a newsletter's readers, a podcast audience, a hobby club.
You care about the archive. You want what you and your people write to still be there in ten years, on a domain you own.
You want to set the rules. Who gets in, what gets ranked, what counts as spam.
It's a bad fit if:
You want reach fast. A new site starts with nobody in it. I call this the empty-room problem, and it is the real cost of owning the room. Chirpper gets roughly 50 real human visitors a day right now. That's after eight months.
You don't want to moderate. Owning the room means owning the mess.
The traps nobody mentions
1. The empty room. Nobody joins a place that looks dead. The fix that's worked best for me: put all of my own writing on my site first, then syndicate copies to dev.to, Hashnode, LinkedIn, Reddit and Hacker News, each linking back home. The borrowed audiences feed the owned one. It's slow, and it takes hours per article, which is why I'm building tools to automate it.
2. You're now a host, with legal duties. I'm not a lawyer, but two US basics every site owner should know: register a DMCA agent with the Copyright Office ($6, renewed every three years) to keep your safe harbor for user uploads, and know that 18 U.S.C. Β§ 2258A requires providers to report child sexual abuse material to NCMEC's CyberTipline once they know about it. If you allow uploads, plan for both before launch.
4. Building is fun, marketing isn't. This is the one that hits me hard everyday. AI made the building fast. It did nothing for getting people to show up. If you won't spend time inviting people in, the site stays empty.
How to start this weekend
Pick one group you already belong to. Don't build for everyone. Build for the 20 people who already talk to you somewhere else.
Write the first prompt like a problem, not a spec. Mine was a question about a social site with no login. Let the AI argue with you before it writes code.
Keep version one tiny. Reading should be free. Decide what it takes to post, and make sure people can reply. Skip profiles, search and notifications at first. You can add them when someone asks.
Pick cheap, boring hosting. A $5 a month serverless plan plus a SQLite database is enough for a long time.
Move your own writing home first. You are the founding content. Then syndicate it out with links back.
Before launch: a DMCA agent, a report button, and a plan for fake accounts.
Build your own, I'll show you how
That's the point of this series. I'm writing up every decision from the Chirpper build, including the ones I got wrong, so you can skip them. If you build your own site, tell me. I'd like to link to it.
Coming up in this series: the login comes back, the architecture I used, hiring my first help, and opening it up for critique.
Richard Kersey, I'm the founder of Chirpper.
Originally published by Dev.to AI. Aggregated on AIWithGhost for educational purposes β full credit and traffic to the original publisher.