About Edit PDF
Editing a PDF feels like it should be straightforward. You click on the text and start typing. In practice, most PDF readers refuse to let you change anything, and the editors that do let you usually cost upwards of fifteen dollars a month.
Edit PDF gives you a free editor that works in the browser. Two modes: Annotate, for adding text, images, shapes, drawings, and highlights on top of the page; and Edit, for changing the actual text that's already in the PDF, with the original embedded fonts preserved so the output looks like the source.
Annotate mode
This is the overlay mode. You're not changing what's in the PDF, you're adding things on top. Text boxes, signatures, arrows, highlights, photos. Each annotation can be moved, resized, rotated, layered, and styled.
The toolbar has six core tools. Hand for panning. Text for typing new content anywhere on the page. Pen for freehand drawing. A shapes menu with rectangle, ellipse, line, and arrow. A highlight tool that drops a translucent yellow rectangle for marking important passages. And an image button that lets you drop in a PNG or JPG (handy for signatures saved as transparent PNGs).
Page navigation is a thumbnail rail on the left. Click a thumbnail to jump to that page. The edits you make on each page stay attached to that page; switching between pages preserves your work.
Edit mode
Edit mode is the harder feature. When you click on a text span in the PDF, the tool replaces that span with an editable text box, then sends your changes to a server-side engine that does the real work. The engine uses PyMuPDF to redact the original text and redraw the new text using the original PDF's embedded font, the original text colour, and a background colour sampled from the actual page so the patch blends in.
The result is a PDF that looks like the original except for the words you changed. Same font, same baseline, same colours, same background. The recipient can't tell from looking that the document was edited unless they have a copy of the original to compare against.
The engine also lets you change the font size, weight (bold or regular), italic style, and colour of an existing text span. So you can fix a typo, change a number, or reformat a heading without redrawing the page.
When you'd actually edit a PDF
Filling in a form that was sent as a flat PDF instead of a fillable one. Click each blank, type your answer, save.
Fixing a typo in a contract before sending it out. Catching a misspelt name or a wrong date saves the embarrassment of a re-circulation.
Updating a date or version number on an old PDF you want to reuse without remaking it from scratch.
Adding a signature, initials, or stamp before forwarding a document.
Marking up a draft. Highlight passages, draw arrows, scribble notes in the margins, then send the annotated PDF back to the author.
Honest limits
Edit mode is best for short text changes. The font matching works because the engine reuses the original embedded font, but if the original PDF had subsetted fonts (which is common), only the characters that were already in the PDF are guaranteed to render. So changing "5,000" to "5,500" works fine. Adding a Greek letter to a document that didn't have Greek text won't.
Saving back to PDF works for text edits, shapes, and images. The output is a real PDF. Pages with annotations get rasterised at 2x resolution and embedded as images to bake in the annotations; pages you didn't touch are copied through losslessly.
Very large PDFs, say more than 50 pages, take longer to analyse on the server. Plan for a few seconds of wait when you switch to Edit mode the first time.