Why I Built My Own Blogging Platform Instead of Using WordPress

I work in cloud security, and most of my day goes into finding what's misconfigured in other people's systems. So when I decided to launch a blog, the question wasn't only where to publish. It was what I was willing to run and defend.

Why not WordPress?
WordPress is the obvious answer, and it's a good one for many people. But its plugin ecosystem is a steady source of vulnerabilities, and I didn't want my first project to depend on patching someone else's code. I also wanted more than a static site: users who could register, write their own posts, comment, and like. That meant a real backend, not a generator with a comment widget bolted on.
The constraints that shaped it

My hosting supports PHP, so the stack was PHP with a SQL database. My usual choice, Spring Boot, wasn't an option.
I wanted nothing installed on my laptop, so everything ran and was tested in Docker containers.
I wanted it to feel like Medium: clean writing, minimal friction, a reader-first layout.

How I built it
I built the platform with AI assistance: [Replit Agent first / Antigravity with Gemini], with me setting direction and reviewing what came back. I started with the media and asset list (logo, images), then the core features: registration, posting, comments and likes. [Add one specific thing that surprised you, good or bad.]

What shipped

User registration and authoring
Comments and likes
An admin portal for user counts and status, because I couldn't query the database directly
Bulk article upload via spreadsheet

What I'm learning
Getting a site live is the easy part. Getting found is the hard part: SEO, indexing, and keeping the production database clean of test data.