People expect a PDF to behave like a Word document and are surprised when it does not. Change one word in a paragraph and the rest of the line refuses to move up neatly. The reason is in the format itself, and once you understand it the limitations stop feeling arbitrary.
A PDF describes a page, not a document
A word processor stores your text as a stream of words with rules about how to lay them out. When you delete a sentence, it re-flows everything after it because it knows where paragraphs begin and end.
A PDF stores the finished result instead: a set of drawing instructions saying which glyph to place at which coordinate, in which font, at which size. There is often no record that a run of glyphs forms a paragraph, or even a word. The file says "draw an H here, an e slightly to the right, an l after that" and nothing more.
That is exactly why PDFs look identical everywhere, which is the whole point of the format. It is also why they resist editing.
What this means in practice
- Text does not re-flow. Replace a short word with a longer one and it will overrun whatever sits to its right, because nothing tells the file where the line should break.
- The font may not be available. PDFs embed subsets of fonts, sometimes only the characters actually used. Type a character that was not in the original text and the embedded subset may not contain it.
- Scanned pages have no text at all. A scan is a photograph. There are no glyphs to edit, only pixels, until you run OCR over it.
- Spacing is often positioned rather than typed. Apparent word spacing can come from coordinates instead of space characters, so extracted text sometimes arrives with odd gaps.
Overlay editing: what browser tools do
Rather than rewriting the page description, browser-based editors draw on top of it. You add a text box, an image, a highlight or a signature as a new layer above the existing content. The original instructions stay untouched.
This handles the majority of real tasks: filling a form that has no form fields, signing a contract, redacting a block, adding a note, stamping a date. It is honest about what it is doing, and nothing about the underlying page degrades.
Where it falls short is genuine copy editing. If you need to rewrite a paragraph and have the text re-flow properly, no overlay tool will do it well, and neither will most desktop editors. Go back to the source document, change it there, and export a new PDF.
Covering text is not redaction
This one matters. Drawing a black rectangle over sensitive text hides it visually but leaves the text underneath in the file. Anyone can select it, copy it, or extract it with a script. Court filings and corporate documents have leaked this way repeatedly.
True redaction removes the underlying content. If you are hiding something genuinely sensitive, the safest approach with a browser tool is to cover it, then flatten the page to an image, accepting that the text becomes unselectable everywhere on that page.
A practical order of operations
- If you still have the source file, edit that and re-export. Everything below is a workaround for when you do not.
- Check whether the text is real text: try selecting it. If you cannot, it is a scan and needs OCR first.
- For additions, signatures and form filling, overlay editing is the right tool.
- For structural changes, consider splitting out the pages you need, replacing one page entirely, and merging the result.
- Keep the original. Every edit produces a new file, and you may want to start over.