Splitting a PDF is one of the most common document-handling tasks, and the right approach depends on whether you want one subset of pages, all pages as separate files, or several named groups. This tool implements all three modes (Extract Pages, Split All, Split by Ranges) entirely in the browser using pdf-lib, which means your files never leave your device. The sections below explain when to use each mode, how to handle edge cases like password-protected PDFs, and how the browser-based approach compares to cloud services for privacy-sensitive documents.
Extract Pages vs Split All vs Split by Ranges
The three modes handle different real-world splitting tasks, and choosing the right one makes the workflow faster. Extract Pages is for pulling out a specific subset: you want pages 1-3, 5, and 7-9 from a larger document as a single new PDF. This is the right mode for taking one chapter out of a book, one section out of a report, or a few signature pages out of a contract. The output is a single PDF with the requested pages in their original order. Split All is for breaking every page into its own file: useful when a single scanned PDF actually contains many logically separate documents (say, 25 scanned tax forms that happened to be in one multi-page file). The output is a ZIP with one single-page PDF per source page. Split by Ranges is the middle ground: you want to split into named groups, where each group may contain several pages. A typical use case is a 200-page report that needs to be delivered as 8 chapter PDFs — you define ranges like "Chapter 1: 1-24, Chapter 2: 25-48, ..." and get 8 named PDFs in a ZIP. Most tools only implement the first two modes, but Split by Ranges is the most practical for document-organization workflows where the source file has natural logical groupings.
Privacy and Why Browser-Based Splitting Matters
Most online PDF-split services upload your files to their servers, perform the split there, and send the result back. This workflow has three issues for anyone handling sensitive documents: the file touches a server you don't control (meaning the operator's security practices and their cloud provider's security practices both affect your data); the file typically sits in that server's storage for 24 hours or longer per the operator's retention policy before deletion; and any security breach at the service exposes your file retroactively. For contracts, medical records, tax returns, legal documents, employee paperwork, or anything else under confidentiality obligations, this cloud-upload model is genuinely risky. This tool uses pdf-lib to perform the entire split locally in your browser's memory — your PDF is read into RAM, manipulated there, and the output is assembled in RAM before the Download button writes it to your local disk. Nothing is uploaded anywhere. You can verify this by disconnecting from the internet after the tool page loads: the tool continues to work perfectly because it has no network dependency beyond the initial page load. For any sensitive document, this local-only workflow is the correct choice. The only limitation is RAM: very large files (500+ MB) may slow down or run out of memory on older devices, but typical contracts, reports, scans, and ebook files split instantly.
Handling Edge Cases and Common Mistakes
A few edge cases come up often enough to warrant specific guidance. Password-protected PDFs cannot be read by pdf-lib or by any browser-based PDF library, because the encryption requires the password to decrypt. You must unlock the PDF first using Adobe Acrobat, Preview on macOS, or a dedicated password-removal tool; once unlocked, the decrypted version works fine here. Very large files (100+ MB) may take noticeable time to load and split — this is memory-bound rather than CPU-bound, so a device with more RAM performs better. Corrupted PDFs fail to open with a clear error message; there is no recovery step in the tool, so you need to use a PDF repair utility on the source before splitting. One common mistake: the Extract mode page count is relative to the original document, not to any subset. If your source has pages 1-100 and you extract pages 10-20, the output PDF has pages numbered 1-11 within itself — not 10-20 as originally numbered. If the original page numbering matters for downstream use (say, legal citations referencing specific page numbers), remember to annotate that mapping before sharing the extract. Finally: splitting does not re-encode content, so images, fonts, form fields, and vector graphics are preserved bit-for-bit from the source. The output pages are exactly as they appeared in the original PDF, just repackaged in a smaller container.