Next.js + MDX: building a blog with no CMS

This post is both the announcement and the test of the exact feature it describes: the blog's content pipeline. No CMS, no database — every post is a .mdx file versioned right alongside the rest of the code.

How it works

Each post lives at src/content/blog/{lang}/{slug}.mdx, with its metadata (title, summary, date) exported at the top of the file as a plain JavaScript object:

export const metadata = { title: "My post", summary: "A short summary", date: "2026-08-15", };

Regular Markdown content goes here.

On the listing page (/blog), I read the available slugs with fs.readdirSync and import each .mdx file to read its exported metadata. On the post page (/blog/[slug]), that same dynamic import also gives me the compiled React component — a single import resolves both content and metadata.

Why this approach

• No CMS: editing a post means editing a file, with real git diffs and history.
• No database: the site stays simple to host — content is just another file in the repo.
• Real Markdown: lists, links, inline code, code blocks, and blockquotes all work with no workarounds.

The look of each element (headings, links, code blocks) comes from src/mdx-components.tsx, which maps Markdown's HTML tags to the rest of the site's color and typography tokens — so a blog post never feels visually disconnected from the rest of the page.

If you're reading this rendered with the right colors, working links, and the code block above with a shaded background, the integration passed the test.

---

Note: this post was AI-generated as sample content, just to test the blog's Markdown/MDX pipeline — it wasn't written by me.