Project structure · imba-pages docs

my-site/
├─ src/
│  ├─ site.imba          site settings: url, default meta, <head> extras, global css
│  ├─ pages/             one file per URL (.imba or .md)
│  ├─ layouts/           default.imba wraps every page; others can be picked by name
│  ├─ components/        your own tags (no special meaning)
│  ├─ islands/           tags that also run in the browser
│  └─ content/           collections (folders) and data modules (files)
├─ public/               copied as-is to the site root: images, icons, robots.txt
└─ package.json

Special names

Path	Meaning
`src/pages/`	Every `.imba` and `.md` file is a page. See Routing
`src/pages/404.imba`	The not-found page (also written to `dist/404.html`)
`src/layouts/default.imba`	Wraps every page, unless a page says otherwise
`src/islands/`	Every tag exported from here is an island
`src/content/<folder>/`	A collection named after the folder
`src/content/config.imba`	Optional checks and sort order for collections
`src/site.imba`	Site-wide settings. See SEO and head
`public/`	Served from the site root: `public/images/a.png` is `/images/a.png`

Names starting with _ are ignored in src/pages and src/content, so _drafts/ is a safe place for work in progress.

Everything else is up to you: src/components/, a src/data.imba file, a src/email.imba helper. Import them like any Imba module.

Generated folders

Both are safe to delete, and belong in .gitignore:

.pages/: files imba-pages generates (the server entry, compiled Markdown, the server bundle).
dist/: the built site.

What happens when it runs

imba-pages scans src/pages, src/islands and src/content, compiles Markdown to HTML, and writes .pages/entry.imba, which imports every page.
Imba's bundler builds that into a server, plus a small browser script per island.
dev runs the server; build runs it once, fetches every page and writes the HTML to dist/.