All processing runs in your browser — no files or inputs are uploaded to a server.
Drop a PDF onto the tool and the page count appears. Type page ranges in the input — `1-3, 7, 10-12` extracts pages 1, 2, 3, 7, 10, 11, 12 into a single output file; `1-3 | 4-6 | 7-9` (pipe-separated) produces three separate files, one per range. The output PDFs preserve the original page rotation, embedded fonts, vector content, and document metadata (title, author, creation date) — the split is structural, not a re-render.
Reach for this when you need only part of a multi-page document — a single receipt from a monthly statement, a chapter from a longer manual, the signature page from a contract. The entire operation runs in the browser via `pdf-lib`; the file never uploads, which matters for financial statements, internal reports, or anything else that should not visit a third-party server. Resulting files are typically the same size as the relevant pages of the input plus some PDF overhead — splitting does not compress content.
file: statement-2026-03.pdf (12 pages) range: 5
statement-2026-03-pages-5.pdf (1 page)
A single-page extract is the most common use — pull one receipt or one transaction page from a longer statement to send to a colleague or expense system. The output file size will be roughly the size of that single page (commonly 50–200 KB for a text-based statement page), not the full 12-page file.
file: api-handbook.pdf (180 pages) range: 1-12 | 13-45 | 46-89 | 90-140 | 141-180
api-handbook-pages-1-12.pdf (12 pages — table of contents + intro) api-handbook-pages-13-45.pdf (33 pages — chapter 1) api-handbook-pages-46-89.pdf (44 pages — chapter 2) api-handbook-pages-90-140.pdf (51 pages — chapter 3) api-handbook-pages-141-180.pdf (40 pages — appendix)
Pipe-separated ranges (`|`) produce one file per range, which the browser downloads as separate PDFs (or as a zip if the tool offers that option). Useful for breaking a long reference document into chapters that can be linked individually or hosted separately. Internal cross-references (`see page 92`) survive the split as text but no longer resolve — those become orphan references.
file: scan.pdf (40 pages) range: 1, 3, 5, 7, 9, 11, 13, 15, 17, 19, 21, 23, 25, 27, 29, 31, 33, 35, 37, 39
scan-odd-pages.pdf (20 pages, kept in original order)
Front-and-back scans where the duplex scanner produced odd pages as fronts and even pages as backs in alternating order can be split out — sometimes one side is what you actually want (signatures, certifications). Type page numbers comma-separated; the tool preserves the order you list them in, so `5, 1, 3` would put page 5 first, then 1, then 3.
No — the operation is structural, not a re-render. `pdf-lib` reads the requested pages, copies their content streams (text, vectors, embedded images) verbatim into a new PDF, and writes the result. Image resolution, font fidelity, and vector precision are preserved exactly. The only changes are the document-level metadata (a new modification date, a slightly different file ID) and the page count itself. If you compare a single page from the split file to the same page in the original, they should render pixel-identically.
No hard cap, but the entire PDF is loaded into browser memory. Practical limit on most laptops is around 200–500 MB; beyond that the tab may freeze or run out of memory. For 1+ GB scans use a server-side tool (`qpdf`, Adobe Acrobat, `mutool extract`) where the file streams through processing instead of sitting in RAM. The pages-loaded threshold is similar — 10,000-page documents stress the parser even if file size is moderate.
Partially. Bookmarks (the outline / TOC in the PDF sidebar) often reference page numbers that no longer exist after splitting; `pdf-lib` keeps the bookmarks but their targets may break. Form fields embedded in the kept pages survive with their values intact; fields in dropped pages are gone. Digital signatures invalidate on any structural modification — a split file is no longer "the same document" the signature was placed on, and Acrobat will mark the signature as broken. If you need to extract pages from a signed document while preserving the signature's validity, you cannot — re-sign the extracted output.
Only if you provide the password to unlock it first. PDFs use two passwords — the **user password** (required to open) and the **owner password** (required to modify, print, copy). `pdf-lib` cannot crack either; it can read encrypted PDFs only if you supply the user password. For PDFs where you have the user password but the owner password is preventing modification, some tools strip the owner password (it is enforced by the viewer, not by encryption) but this tool does not — out of respect for the document author's explicit restriction.
They are inverse operations. Split takes one PDF and produces multiple (or one smaller) PDFs. Merge takes multiple PDFs and produces one combined PDF. The PDF Merge tool on this site is the counterpart; the two together cover the most common page-level PDF restructuring. Round-trip is lossless when no re-encryption or re-signing is involved — split a 50-page PDF into 5 chunks of 10, then merge them back, and the result is structurally identical to the original (same content streams, same fonts, same images), differing only in document metadata.
PDFs often embed fonts (used across all pages) and a shared object pool (images referenced from multiple pages, color profiles, vector libraries). When you split out a single page, `pdf-lib` must include all the resources that page references — even if those resources were also referenced by other pages, since they cannot be shared across files. So extracting page 5 of a 12-page document might include the same fonts and images as the full document. For genuine size reduction, run the output through a PDF optimizer (`Ghostscript`, `qpdf --object-streams=generate`, online optimizers). The split tool focuses on correctness, not compression.
PDF (Portable Document Format, Adobe 1993, ISO 32000 since 2008) is a page-description format that wraps fonts, vector commands, raster images, and document metadata into a single file. The container is a graph of *indirect objects* — fonts referenced once, used everywhere; images stored once, referenced from any page that includes them. Page-level operations like splitting and merging walk this graph and produce new files by selecting subsets of pages and the resources they require.
**Splitting** is mechanically simple: identify the pages to keep, traverse their content streams to find which fonts/images/color spaces they reference, copy those resources into the new file, and write the page tree referencing them. The hard part is doing this without breaking anything — bookmarks, named destinations, hyperlinks, and form fields can reference page numbers (`/Page 5`) or object IDs that need updating in the new file. Robust PDF tools (`pdf-lib`, `qpdf`, Adobe Acrobat) handle the common cases; edge cases like XFA forms, JavaScript actions, or encrypted file attachments often require commercial tools.
Three adjacent concepts shape the ecosystem. **PDF/A** (ISO 19005) is a constrained PDF subset for long-term archival — fonts must be embedded, encryption forbidden, JavaScript banned. Splitting a PDF/A document risks dropping fonts that the rest of the document depended on, breaking PDF/A compliance; the split tool does not preserve the PDF/A claim. **OCR (Optical Character Recognition)** turns scanned image pages into searchable text — a separate operation from splitting, but often needed before splitting if the document came from a scanner. **PDF/UA** (Universal Accessibility) tags PDFs for screen readers and adds reading-order metadata; splitting preserves the tags on kept pages but can leave dangling cross-references. For accessibility-critical documents, re-validate the output with PAC (PDF Accessibility Checker) after splitting.
Combine multiple PDFs into a single document. Reorder pages, then download — files never leave your browser.
Convert each PDF page to a PNG or JPEG image. Adjustable scale, range selection, multi-page ZIP download.
Shuffle a list of names into balanced teams. Choose either the number of teams or the size per team; teams come out within 1 member of each other. Optional team names and one-click copy.
Create vector-PDF quotes, bills, and invoices for small businesses. Switch document type with one click, fill line items, and download a clean A4 PDF — Korean / Japanese / English templates with locale-aware tax (VAT 10%, 消費税 10%), currency, and business-number formats.
