Dev.to WebDev 🛠 Dev 👁 0 📖 2 min read

A CMS without a database: the kinds there are and when each fits

Whoever looks for a CMS without a database usually wants three things: less maintenance, less attack surface and hosting that runs anywhere. But the term covers four very different designs. Which one fits depends on who

Whoever looks for a CMS without a database usually wants three things: less maintenance, less attack surface and hosting that runs anywhere. But the term covers four very different designs. Which one fits depends on who maintains the content and who looks after the system.

1. Flat file CMS

Systems such as Kirby or Grav run with PHP on the server and store content in files instead of a database. The client gets a full backend, and a backup is a copy of the folder.

The database is gone, the software on the server stays. It needs updates like any other CMS, and the server generates the website on every request. Kirby needs a licence per website, Grav is open source.

2. Git based CMS

Decap, Sveltia, TinaCMS or Pages CMS are interfaces for files in a Git repository. The client changes a field, the CMS writes the change into the repository, and a generator such as Astro, Hugo or Eleventy builds the website from it.

After that only finished HTML lies on the web server. In return you need a generator, Git hosting, a sign in for the client and a configuration that describes every field. For clients without technical experience the interface is often bare.

3. Headless CMS

With Sanity, Contentful or Storyblok the content lives with the provider. Your website has no database; the service of course has one. You just do not run it.

The editor is mature. You build and host the website for it separately, and it has to match the data model. That pays off when the same content appears in several places, such as the website and an app.

4. A platform with static delivery

Editing happens in the platform, and finished HTML is delivered. The content lives in the platform, not on the server that delivers the website. Visitors reach neither a login nor a CMS, and there are no plugins anyone would have to update.

On the newer ones the website is built by your AI agent, for example Claude Code. It creates components with fixed fields, and these fields become the editor for the client. Hosting, forms, approvals and a preview behind a password are included.

The four kinds compared

Where does the content live? What runs on the web server? Who installs updates? Editor for clients
Flat file CMS files on the server PHP and the CMS you full
Git based CMS files in the repository only HTML you, for generator and CMS simple
Headless CMS with the provider whatever you build the provider for the CMS, you for the website mature
Platform with static delivery in the platform only HTML nobody fits the website exactly

When each kind fits

  • You want everything on your own server and know PHP: a flat file CMS.
  • You work with Git and a generator anyway, and the client changes little: a Git based CMS.
  • The same content goes to several channels: a headless CMS.
  • You look after many client websites, an AI builds them, and the clients maintain them: a platform with static delivery.

This article was written by an AI agent and checked by the team at Cleverpage, which builds a platform of the last kind described here. It first appeared, with a section on our own product, at cleverpage.ai.

📰 Read the original article on Dev.to WebDev

Originally published by Dev.to WebDev. Aggregated on AIWithGhost for educational purposes — full credit and traffic to the original publisher.