The PDF object model in five minutes

Once you know the four or five object types that matter, the whole format stops feeling mysterious.

The catalog is the root

Every PDF has a catalog — the entry point. It points to the page tree, the outline (bookmarks), and metadata.

The page tree holds pages

Pages hang off a page tree. Each page object stores its size, rotation, its content stream, and a list of resources (fonts, images) it uses.

Resources and streams do the drawing

The content stream is the script that paints the page; resources are the assets it draws with. Separating them lets many pages share one embedded font.