Invert PDF colors (dark mode)

No, and that is the nice part. A full-page white rectangle is drawn over each page using the Difference blend mode, which computes one minus the backdrop for every colour channel. That is an exact inversion done inside the PDF's own graphics model. Text stays text, selectable and searchable, and images keep their original data.

Barely. What is added per page is one rectangle and a blend mode instruction, so the growth is a few bytes rather than a re-encoded page. Nothing is resampled or recompressed, which is why this stays fast on a long document.

Run it again. Inverting an inversion returns every channel to where it started, so a second pass through this tool undoes the first. Your source file is never modified either way; the result is a new document with an -inverted suffix.

Because the effect lives in the file as a blend mode instruction rather than as baked pixels. What you see depends on your reader applying that instruction. Blend modes are a standard part of the PDF format, so mainstream readers handle it, but a simple or embedded viewer might not.

None. It inverts every page of every file you drop, with nothing to configure. There is no partial inversion, no way to keep images upright in colour while flipping the text, and no page range. If you only want some pages inverted, extract them first, invert, then merge.

No. pdf-lib loads the file, draws the rectangles and saves it, all in a Web Worker in this tab. Drop as many PDFs as you like and each comes back as its own inverted file. Nothing is sent anywhere and the tool runs with your network off.