A file-based markdown engine. Point it at a folder of markdown files — an Obsidian vault, for example — and get a fast three-column news site: menu, feed, reader. No database, no logins, no users.
📁 File-based architecture — organize articles in folders.
✍️ Markdown-first — write and manage content using simple Markdown files.
🗂️ Automatic article discovery and organization.
🌐 Publish and browse your news through a web interface.
⚡ Lightweight and easy to maintain.
🔧 Simple content management with Git-friendly files.
Live preview · https://miloser.dev/newspaper/
Fork & Try · https://github.com/miloserdev/newspaper

Top comments (3)
A documentation engine earns trust when publishing is separate from parsing. I’d version the source content, validate links and front matter in CI, generate a preview artifact, and promote an immutable build rather than rendering arbitrary live markdown at request time. That gives authors a fast workflow while keeping a broken document from becoming a production incident.
For a file-based Markdown publisher, I’d validate front matter and relative links before generating the three-column view, then build a preview from the exact Git commit being reviewed. That catches missing assets or malformed metadata before publication without making the live reader experience depend on parsing whatever happens to be in the folder. Keeping a few representative vaults as fixtures also helps catch path and nesting edge cases.
Repo updated, thanks a lot for the feedback!