How to Edit a PDF When You Don’t Have the Original: What Works, What Breaks, and How to Fix It

How to edit a PDF: the short answer

Choose the method by what you need to change. To add a note, fill in a form or sign, use the free viewer you already have: Edge, Chrome, Preview on a Mac or Markup on an iPhone. To reorder, remove or rotate pages, use a page tool, which leaves the text untouched. To rewrite text, convert the PDF to Word, make the change, check five things, and save a new PDF. For a scan, run text recognition first. When the change must be perfect, ask for the original file.

The rest of this guide explains why PDFs resist editing, gives a table that matches each kind of change to the method that suits it, walks through each method from the lightest to the heaviest, and ends with a worked example: a one-page resume that existed only as a PDF, converted, fixed in Word in twelve edits and saved as a new PDF with every word of the original in place.

Why a PDF resists editing

A PDF is built to look the same everywhere, and it achieves that by recording the page rather than the document behind it. Its page content is a series of drawing instructions: put these characters at this position, in this font, at this size, then draw this line, then place this picture. The format describes itself as “a complete description of a fixed-layout document”, in the words Wikipedia’s summary of the standard uses (https://en.wikipedia.org/wiki/PDF). What it usually does not record is the thing a word processor works with: that these lines form one paragraph, that this line is a heading, that these numbers sit in the cells of a table.

Three consequences follow. First, changing one word does not make the rest of the line move over, because there is no line to reflow, only characters at fixed positions. Second, the embedded font is often a subset holding only the characters the document used, so a letter that never appeared in the original is not there to type. Third, some PDFs carry optional structure tags that describe headings and paragraphs for screen readers, but many do not, and editing tools cannot count on them. That is why a “PDF editor” either works on small patches of text in place or rebuilds the whole page as a document first.

A PDF can also be locked. A password to open it, or restrictions set by its author on copying or changing it, stop editing tools on purpose. A tool that respects the file tells you so rather than working around it; the honest fix is to ask whoever sent it for a version without the restriction.

Which method for which change

Most people who search for how to edit a PDF need one of seven changes. Each has a method that does the job with the least disturbance to the rest of the file:

If the file is too big to send once you have edited it, that is a separate job: How to reduce PDF file size covers where the size comes from, and How to send a large PDF covers attachment limits.

Route 1: annotate, fill in and sign in a free viewer

The lightest edits never change the original content. They add a layer on top of the page: a highlight, a comment, a line of typed text, a drawn signature, or the answers in a form’s fields. Every major platform now includes a viewer that does this for free, so there is nothing to install and nothing to upload.

• Microsoft Edge (Windows and Mac): Microsoft’s documentation for the built-in PDF reader says you can annotate files with ink and highlighting, add text notes, and fill in basic forms; forms that rely on JavaScript and XFA forms are not supported (https://learn.microsoft.com/en-us/deployedge/microsoft-edge-pdf). The same reader can check certificate-based digital signatures.
• Google Chrome on a computer: Google’s help page lists highlighting, annotating, drawing, adding text, filling in forms and signing in Chrome’s PDF viewer (https://support.google.com/chrome/answer/16215622). On a Chromebook, Google points to the Gallery app for text annotations instead.
• Preview on a Mac: Apple’s guide says you can type into form fields, add a text box where a form will not accept typing, and create a signature with the trackpad, the camera or an iPhone (https://support.apple.com/guide/preview/fill-out-and-sign-pdf-forms-prvw35725/mac). Preview adds text and signatures; it does not rewrite the text already on the page.
• iPhone and iPad: Markup draws, adds text, shapes and a signature on PDFs from the Files app or Mail (https://support.apple.com/en-us/119875).

A typed text box can cover a wrong date or a misspelt name and look corrected on screen and on paper. Be clear about what that does and does not do: the original text is still in the file underneath and can be selected, searched or revealed by removing the box. That is fine for correcting your own form; it is not redaction, and it is not a way to hide information from anyone who receives the file.

Route 2: change the pages, not the text

Many “edit” jobs are really about pages: a scanner slipped in a blank page, a contract needs its signed page added, an appendix is out of order, or a phone scan came out sideways. Page tools handle these without converting anything, so the text, fonts, links and any bookmarks on the pages you keep stay exactly as they were.

• Take pages out, keep only some, or split a file into parts with Split PDF. Deleting a page removes it from the file; it does not change the pages around it.
• Put files together, reorder pages across them and insert blank pages with Merge PDF.
• Turn sideways pages upright with Rotate PDF, which changes the page’s rotation setting rather than redrawing it.
• Make an edited file smaller before sending it with Compress PDF.

All four run in the browser, so the document stays on your device. Are online PDF converters safe? explains how to check that for any site in about a minute, using the Network tab in your browser’s developer tools.

Route 3: convert to Word, edit, save a new PDF

When the words themselves have to change, the dependable method is to turn the PDF back into a document, edit it where editing is easy, and make a fresh PDF. Microsoft Word can open a PDF directly and convert it; Microsoft’s support page says the conversion “works best with files that are mostly text”, such as business, legal or scientific documents, and that the original PDF stays as it was (https://support.microsoft.com/en-us/word/opening-pdfs-in-word). The same page lists what does not convert well: tables with cell spacing, page colours and page borders, tracked changes, frames, footnotes that span more than one page, endnotes, audio, video and PDF active elements, PDF bookmarks, tags and comments, and font effects such as glow or shadow.

Without Word, or when you would rather not send the file anywhere, PDF to Word does the conversion in your browser. It offers two kinds of result. Editable text rebuilds real paragraphs, headings, lists and tables that reflow as you type, which is what you want for rewriting. Exact layout places every block where it was on the page, which keeps the look of a form or a designed page but is harder to edit. It finishes with a report of what it recognised and what to check, and you can switch a single badly recognised page to exact layout.

Google Docs can open a PDF too: upload it to Google Drive and open it with Google Docs. Google’s help page says that route uses text recognition, works best on files of 2 MB or less with text at least 10 pixels high, and that lists, tables, columns, footnotes and endnotes are not likely to be detected (https://support.google.com/drive/answer/176692). It is handy for pulling the text out of a short document, less so for keeping its structure.

The five checks after converting

Whichever converter you use, the same five things go wrong most often. Checking them in this order takes a few minutes and turns a converted draft into a clean document:

• Headings. Open View, Navigation Pane in Word. Every section title should be listed. Where one is missing, click it and apply Heading 1, 2 or 3 from the Styles gallery. Headings are what make a table of contents and the bookmarks in the new PDF work.
• Tables. Click inside each table and read it row by row against the PDF: the right text in the right cell, no two cells run together, no row split in two. A table that came across as lines of text with tabs can be turned back into a table with Insert, Table, Convert Text to Table.
• Page breaks. The text flows again in Word with Word’s spacing, so a heading can end up alone at the foot of a page and a document can gain or lose a page. Delete stray empty paragraphs, set Keep with next on headings (Paragraph, Line and Page Breaks), and adjust margins last.
• Headers and footers. Double-click the top or bottom of a page. Page numbers should be fields that count up (they show shaded when clicked), not the same typed number on every page, and a different first page should be set if the original had one.
• Hyphenation and line ends. Words hyphenated at the end of a line in the PDF can arrive as two halves, and some lines can end in a hard break. Search with Find for a hyphen followed by a space, and use Replace with ^l (a manual line break) to find breaks inside paragraphs.

Then save a new PDF from Word with File, Save As or Export, and PDF as the type. How to save a Word document as a PDF goes through the options that matter, such as bookmarks from headings and document properties, on Windows, Mac, phones and Google Docs. If Word is not installed, Word to PDF makes the PDF in your browser with the fonts embedded and the links working.

Route 4: scanned PDFs need text recognition first

A scanned PDF is a photograph of each page. To tell, try to select a single word in your viewer: if you can only draw a box over the page, there is no text to edit, only pixels. Text recognition (OCR) reads the picture and produces text that a converter can then organise into paragraphs. PDF to Word does this on your device for pages that need it, with English, Spanish, French, German, Portuguese and Italian available, and can also return the scan as a searchable PDF with an invisible layer of recognised text, which is all some people need.

Recognition is only as good as the picture. The documentation of Tesseract, the open-source OCR engine, says it works best on images of at least 300 DPI, and lists the other things that hurt accuracy: skewed pages, uneven backgrounds that make it hard to separate text from paper, noise, dark borders and text that is too small (https://tesseract-ocr.github.io/tessdoc/ImproveQuality.html). On three test scans we made at 300 DPI with a slight tilt and speckle, 99.0%, 98.3% and 91.4% of the words were read correctly, the lowest on a page with a coloured table and small print. If you are scanning again anyway, scan at 300 DPI, flat and straight, and proofread names, figures and dates afterwards.

When only the original file will do

Some documents should not be rebuilt at all. A designed brochure, a magazine page or a document full of equations will need more repair after conversion than it would take to recreate from the source. A contract or an application where the exact wording and layout carry weight is safer changed in its original file by the person who made it. And a document someone else signed digitally should not be edited, for the reason in the next section. In all of these, the fastest route is usually an email asking the sender for the Word, Pages or InDesign file, or for a corrected PDF.

What editing does to signatures, and a word on validity

A digital signature in a PDF is bound to the exact content that was signed. The European Commission’s FAQ on electronic signatures describes an advanced electronic signature as linked to the signed data “in such a way that any subsequent change of the data is detectable” (https://ec.europa.eu/digital-building-blocks/sites/spaces/DIGITAL/pages/880312429/eSignature+FAQ). In practice, University of Illinois guidance on signature errors in Acrobat says a signed document must stay byte-for-byte identical, and warns never to run a signed document through a merger, converter or compressor (https://answers.uillinois.edu/ahs/159222). Converting a signed PDF to Word and back produces a new, unsigned document; the signature does not come with it. Whether you may change a document someone else sent you depends on what it is and where you are; when it has legal effect, ask the other party for a corrected version rather than editing it yourself.

Worked example: a resume that exists only as a PDF

We took the one-page sample resume produced by this site’s resume builder, a typical case of a document you might have only as a PDF, and treated it as if the original were lost. It is set in the Inter typeface with spaced-out capital section headings, job titles with dates aligned on the right, bullet lists and a skills section. We converted it with PDF to Word in editable mode, opened the result in Microsoft Word, made the fixes a person would make, saved a new PDF from Word, and compared its words with the original PDF. Converting took under a tenth of a second on a desktop PC.

The conversion report said: 30 paragraphs rebuilt as editable text and 6 headings set as Word headings; 10 list items found and rebuilt as one real Word list; 5 lines of spaced-out capitals kept with a space between letters, because the PDF spaces the letters out and does not say where one word ends; and the fonts Inter and InterTabular set in Arial. Word opened the file without a repair prompt, on one page.

Twelve edits in all, a few minutes’ work. Before the fixes, the converted file held 97.5% of the original PDF’s words in the right order; the rest were the headings broken into single letters. After the fixes, Word’s new PDF contained all 315 words of the original in the same order, fitted on one page, and had six bookmarks built from the headings, one for each section, which the original PDF did not have. The page was not a pixel copy of the original: Word laid the text out with its own line spacing and in Calibri rather than Inter, so lines break in slightly different places and the page needed narrower margins to fit. For a resume that is usually an acceptable trade, because the content and the structure are what an employer’s software reads. For a designed page it may not be, which is when exact layout, or the original file, is the better choice. Keep the converted Word file as your new original, so the next change starts from a document rather than a PDF again.

Frequently asked questions

More on working with PDFs and other files is in the Files and PDF guides, including when a PDF is the right format in the first place in PDF vs JPG vs PNG.