Edit the pages of a PDF

How to use it

1. Choose a PDF. Every page appears as a thumbnail; add more PDFs or photos at any time and their pages join the end.
2. Work on a page with the controls under it: ↺ ↻ to turn it, ⧉ to duplicate, ← → to move it, ✕ to remove it.
3. Tap pages to select several, then act on all of them at once — or save just those to a new file.
4. Press Download. The document is rebuilt from your originals, once, with every change applied.

Every change is an instruction, not a save

Doing this job with single-purpose tools costs a round trip each: rotate page 3 and download, re-open and delete page 7, re-open and append a scan. Each pass writes a new PDF, and a document that has been saved five times carries the residue of five saves — some tools append their changes rather than rewriting, so the file grows while the visible page count does not.

Here nothing is written until the end. Your edits are held as an ordered list of instructions — this page of that source, turned this far — and the file is assembled once from the original bytes when you press Download. Undo is therefore exact rather than approximate: it steps the list back, and no partly-saved document exists to recover from.

Pages are copied, not re-rendered — and what that rules out

A page is moved into the new document as the object it already is: its content stream, its embedded font subsets and its compressed image data are copied across untouched. Text stays selectable and searchable, links keep working, and a 300 DPI scan does not lose a generation. This is also why turning a page is instant on a large document — rotation is a single number in the page dictionary, not a re-render.

It is also the honest limit of this tool. Because pages are copied rather than rebuilt, it changes which pages exist and how they sit, not what is printed on them: there is no editing the text inside a page here. PDF stores text as positioned runs of glyphs against a subset of a font, with no paragraphs to re-flow, so rewriting a sentence means re-measuring the line, re-embedding characters the subset may not contain, and reprinting everything after it. Tools that offer it either rasterise the page or approximate the font, and both show. If you need the words changed, change them in the document this PDF came from.

Questions