# Extract Pages from a PDF and Keep the Page Numbers Readable
You can handle this directly in your browser with our free add page numbers tool — nothing gets uploaded.
Extracting a page range from a multi-page document sounds simple: you pick pages 27 through 34, save the result, and move on. In practice, the operation silently destroys a lot of the information that made the original document usable. Printed folios such as **Page 27 of 40** become meaningless the moment the new file only contains eight physical pages. Bookmarks vanish, filled-in form fields are lost, embedded links stop pointing at the right place, and any digital signature covering the extracted range is voided. This article walks through what happens to each of those elements, how the pieces survive or fail to survive, and how to renumber the pages so that the new document remains readable.
A PDF stores page objects in an ordered array. When you extract pages, the software copies that subset of objects into a new file and renumbers them starting from page 1. The **printed** page number, however, is not stored anywhere the extractor can see: it is usually a piece of text sitting inside the page's content stream, or it is a string in a header or footer template that the layout engine baked into each page during export.
That means two different outcomes are common:
1. The printed folio is a literal text object. It survives the extraction unchanged, so the new page 1 still says *Page 27 of 40*. Readers have no way to map the printed number to the physical position in the new file.
2. The folio was generated by a print-time template that included the total page count of the original. After extraction the template is gone; the original number is a stale artifact.
Neither result is a bug. It is just what extraction is: it copies, it does not re-layout.
The PDF outline tree is a document-level structure, not a page-level one. When you extract pages 27 through 34, every outline entry that did **not** point into that range is dropped. Entries that did point into the range are usually dropped as well, because their destination dictionary references an absolute page index that no longer matches. The result is a new file with no bookmarks at all, even if the source had a deeply nested table of contents.
Some tools attempt a partial rescue: they keep outline items whose destination lands inside the extracted range and rewrite the destination to the new index. Others do not bother, and most free command-line utilities fall in the second camp. If a TOC matters, expect to rebuild it by hand or by generating a fresh outline from headings in the new file.
Form fields (AcroForm) are also document-level objects: a dictionary of field names bound to page-absolute widgets. When the widgets disappear because their host pages were not extracted, the field definition is stripped. Even when the widgets are inside the extracted range, the **values** the user typed in are stored outside the content stream, in a separate field-object tree. Most extractors drop that tree by default, so a filled-out survey or application prints as a blank template. The field structure itself can survive only if the extractor explicitly preserves the AcroForm dictionary and rewrites widget pages to the new indices; that is a niche capability and rarely the default.
Links have two kinds of destinations and their fates differ:
In short, links that go out of the document usually survive; links that jump to a specific page inside the document usually do not.
A digital signature in a PDF is a cryptographic hash over the document's byte range. Any change, even copying out a subset of pages, changes the hash and the signature status. The new file will report its signature as **invalid** or **unverified**, and any previously applied approval will no longer be legally meaningful. If the extracted pages carried an executed signature, plan on having the signatory re-sign the new document.
Once you understand what extraction throws away, the next step is to give the new file a coherent set of numbers. The most flexible approach is to strip the old folio (often a fixed position in the header or footer) and let your output tool render a fresh one.
A typical workflow looks like this:
1. Open the source PDF in a tool that exposes page objects (e.g. a desktop editor with scripting, or a CLI that can add text stamps).
2. Select the page range you want and export it to a new file.
3. Apply a header/footer template to the new file. Use a *physical* counter, not a stored folio, so page 1 of the new file says *Page 1 of N* with N equal to the new total.
4. If the original printed numbers matter to the reader, add a small parenthetical annotation in the footer, for example *(original page 27)*, so cross-references still resolve.
5. Re-add any bookmarks you need, as a new outline built from the headings now visible in the file.
6. If the document was a form, fill it out again, or merge the original field values into the new file via a form-aware tool before finalizing.
7. If it was signed, re-sign after the renumber is locked, because every byte change after that will void the new signature too.
The order in steps 6 and 7 matters: renumber, rebuild fields, then sign last.
The table below summarizes how typical extractors handle each artifact when you take a page range out of a 40-page source. It is a rough generalization; specific tools differ, so verify against the utility you actually use.
NoFilePDF, one of the handful of utilities that exposes a page-level copy primitive, behaves in the middle row of this table: it copies the literal text, rewrites outline entries that land inside the range, and preserves remote links, but it does not carry AcroForm values and it does not touch signatures. NoFilePDF is useful when you want a quick, scriptable extraction without a GUI; treat its output as a starting point, not a finished document.
**Q1: Can I extract pages 27 to 34 and keep the original printed number *Page 27 of 40* on the new pages?**
Yes, and you probably do not want to. The number is a literal string or a baked-in template, and extraction copies it as-is. You can keep it, but it will mislead anyone who only has the new eight-page file. The better practice is to strip the old folio and stamp a new physical counter, optionally with an *original page X* annotation for cross-referencing.
**Q2: Will my filled-out form still work after I extract the pages that contain the fields?**
In almost every common tool, no. The field values live in the AcroForm dictionary, which is not part of the page content stream, and the extractor does not carry it over. You will see the fields present but blank, or the fields will be gone entirely. If the form values matter, fill in the new file again, or use a form-aware tool that explicitly merges values into the extracted copy before you finalize it.
**Q3: Does extracting pages void a digital signature on those pages, and is there any way around it?**
Any byte change, including extraction, changes the document hash and the signature is reported as invalid. There is no way to "carry over" the original signature into the new file. If the pages were signed, the only reliable path is to have the signatory sign the new document after you have finished renumbering, rebuilding fields, and freezing the byte stream. Sign last; signing early and then editing will just invalidate the new signature as well.
Extraction is a copy operation, not a layout operation. The printed numbers, bookmarks, form values, internal links, and signatures that made the source document readable all live outside the page content stream, and they will not follow the pages you keep. Plan to strip the old folio, rebuild a physical header or footer, recreate the outline, re-enter any form data, and re-sign if needed. Do those steps in that order and the extracted range will read like a document that was always meant to stand on its own, rather than a mutilated fragment of a larger one.
Related reading: add blank page to pdf, add page numbers, add page numbers formatting.