Extract PDF Pages

Page extraction from a PDF involves creating a new document that contains a subset of the original's page objects along with all their dependencies. Understanding the PDF object graph makes it clear why this is more nuanced than simply copying page entries from one file to another.

Each page in a PDF is a dictionary object containing entries for its media box (physical dimensions), content stream (drawing instructions), and a resources dictionary that maps names to font objects, image XObjects, graphics state parameters, color spaces, patterns, and shadings. These resource objects are often shared across multiple pages — a corporate document might reference the same logo image and header font on every page, stored once and pointed to by each page's resource dictionary. During extraction, the tool must follow these reference chains to identify and include every object that the extracted pages depend on.

The PDF cross-reference table maps object numbers to byte offsets within the file. When building the extracted document, all copied objects are assigned new object numbers and a new cross-reference table is constructed. Indirect references — pointers from one object to another using the "N 0 R" syntax — must be updated to reflect the new numbering. Any reference to an object that is not included in the extraction (because it belongs to a non-extracted page) must be either resolved or nullified to prevent broken references in the output.

Page labels — the optional labeling scheme that assigns logical names to pages (e.g., "i", "ii", "iii" for front matter, then "1", "2", "3" for body) — are defined in the document catalog's /PageLabels number tree. When extracting a subset of pages, the label assignments must be remapped to the new page indices. If pages 10-15 of the original carry labels "A-1" through "A-6", those labels should appear on pages 1-6 of the extracted document. Incorrect label mapping causes confusion in PDF readers that display page labels in their navigation controls.

The document outline (bookmarks) is stored as a tree of outline items, each pointing to a destination on a specific page. When extracting pages, outline items whose destinations fall within the extracted range should be preserved, while items pointing to excluded pages must be pruned from the tree. If an outline parent node has no remaining children after pruning, it too should be removed to maintain a clean navigation structure. This recursive pruning ensures the output document has a coherent, functional bookmark tree.

Named destinations — a mechanism for cross-referencing specific locations within a PDF — are stored in the document catalog's /Names dictionary under the /Dests key. These destinations enable deep linking from external documents, web pages, and within the document itself. During extraction, named destinations that reference extracted pages are preserved in the output's name tree, while those referencing excluded pages are removed. This ensures that any external links targeting the extracted content continue to function correctly.

Extracting specific pages from a PDF is a core document workflow across many fields:

• Contract excerpts — Legal professionals extract signature pages, specific clauses, or exhibit sections from lengthy contracts to create standalone reference documents for counterparties or internal review.
• Report summaries — Extracting the executive summary and key findings from a 100-page research report creates a concise document for stakeholders who need only the high-level conclusions.
• Portfolio assembly — Designers and architects extract selected project pages from multiple PDF portfolios to assemble a custom presentation tailored to a specific client or competition submission.
• Academic research — Researchers extract relevant chapters or sections from journal articles and textbooks to compile focused reading materials for literature reviews, seminars, or study groups.
• Insurance claims — Claims adjusters extract specific pages from multi-document bundles — a particular invoice, a specific photograph, or a relevant medical report — to attach to individual claim records.
• Immigration and visa applications — Applicants extract specific pages from passports, bank statements, and supporting documents as required by consulate submission guidelines, which often specify exact page requirements.

Extracting pages locally in your browser eliminates the privacy concerns inherent in cloud-based PDF tools. Documents containing personal information, financial data, trade secrets, or attorney-client privileged content never leave your device. This makes the tool compliant with data handling requirements in regulated industries where uploading files to third-party servers is prohibited by policy or law.