What the X-ray measures and how it calculates it
Size breakdown
A PDF is a container of objects. The tool goes through them one by one and adds up the space each takes as stored in the file: images, embedded font programs and page content (text, lines and shapes). Whatever doesn't fit those groups —metadata, color profiles, attachments, internal indexes and leftovers from earlier edits— is grouped under “structure and other”. It's a faithful estimate, but not an exact one: an image inserted inside a page's own content counts as text and vectors.
Fonts
For each font it shows whether it's fully embedded, embedded as a subset (only with the characters the document uses) or not embedded. A non-embedded font takes no space, but each computer replaces it with one of its own and the document may look different; a fully embedded one can add hundreds of kilobytes.
Images and pages
It counts the distinct images the file contains. Transparency masks, which accompany some images, are shown separately so as not to inflate the figure.
Color palette
The first page is drawn on an invisible canvas and its pixels are grouped by color. The result is the five colors covering the most area, with their HEX code ready to copy. If you're only interested in brand colors, you can hide whites, blacks and grays.
Your document never leaves your device
The analysis happens inside the tab, on a separate thread: the PDF is read from your disk into the browser's memory and isn't sent to any server. You can check it as we explain in this guide.
And then what?
If most of the size is images, the compressor can bring the file down to the exact size you've been asked for. If you want to understand why a scan goes over 10 MB, we explain it in why a PDF is so large.