PDF Page Extractor

Page extraction is often the last step before a PDF leaves its original audience. A reviewer may need three exhibits from a long case file, an instructor may need selected worksheets, or an administrator may need a packet assembled in a different order. The new document is useful only when every output page can be traced back to the intended source position.

PDF position and printed page number are not always the same. Covers, Roman-numbered front matter, inserted scans, and custom page labels can make the page marked “1” the third or tenth page in the file. Extraction selectors normally address the file’s one-based positions, so a quick check in a PDF viewer prevents an otherwise plausible but wrong range.

Source position

The first, second, or later page object in the PDF, counted from 1.

Output sequence

The order in which selected source pages appear in the new PDF. It may preserve a custom order and intentional repeats.

Page manifest

A row-by-row record linking each output page to its source page and selector.

Two extraction jobs can name the same source pages and still produce different documents. A typed sequence such as 6, 2, 6 can create a three-page packet with page 6 repeated. A sorted unique sequence turns the same request into pages 2 and 6 once each. Choosing between those outcomes is a document decision, not a formatting preference.

Copying pages also creates a new PDF rather than preserving the source file as a signed whole. Digital signatures can become invalid, and bookmarks, named destinations, forms, page labels, or tagged structure may change or disappear. Extraction is not redaction either: sensitive text, comments, links, metadata, attachments, or hidden content on a copied page still require a separate review.

A reliable handoff therefore checks both the mapping and the opened download. The manifest confirms where each page came from; reopening the generated PDF confirms that the pages render, appear in the intended order, and still meet the document’s legal, accessibility, or records requirements.

How to Use This Tool:

Load one unlocked PDF, express the page sequence by source position, and verify the generated copy before sharing it.

1. Choose Browse PDF, drop one file into the source area, or select Use sample PDF. The file must be a readable PDF no larger than 150 MiB.
2. Enter Pages to extract. Use individual positions, ascending ranges, open-ended ranges, odd, even, or all. Check the warning below the field if a page is outside the document or a range counts backward.
3. Choose Typed order, preserve repeats when order and repeated pages are intentional. Choose Ascending unique source pages when the output should be sorted and duplicates removed.
4. Open Advanced only if the filename or Extraction guard needs changing. The guard accepts 1 to 500 output pages and blocks a selector that expands beyond it.
5. Run Extract selected pages. Compare the output count and Page manifest with the request, download the PDF, then open the downloaded copy for a visual and signature-status check.

Interpreting Results:

The output page count includes repeated source pages in typed-order mode. Source positions counts distinct source pages, while Repeated positions shows how many extra copies the final sequence contains.

• Use Page manifest to confirm each output-to-source mapping, especially after custom ordering or repeated selections.
• Use Source coverage to see how many source pages were included at least once and how many were omitted.
• Do not treat a successful extraction as proof that signatures, bookmarks, forms, tags, labels, or hidden information remained suitable for the intended use.

Technical Details:

Extraction begins with a page-expression grammar, expands that expression against the loaded page count, applies one sequence policy, and copies the resulting zero-based page references into a newly created PDF. The reader-facing selectors remain one-based throughout.

Rule Core:

Transformation Core:

Selectors are read left to right after surrounding spaces are trimmed and letter keywords are lowercased. Commas separate tokens. The expression may contain at most 1,000 characters, and the source PDF may contain at most 100,000 pages.

The extraction guard is applied after sequence handling. A request succeeds only when the final sequence contains at least one page and its length is no greater than the selected guard, which itself must be from 1 through 500.

The new file receives copied page objects plus a new title and producer value. It is saved as a separate PDF with object streams enabled. Document-level structures are not independently reconstructed or certified, which is why the downloaded copy needs review in a full PDF viewer.

Privacy Notes:

The selected PDF is read and rewritten in the current browser tab. Its contents are not uploaded for extraction. The generated PDF and manifest exports are still local files that may contain confidential material, so store and share them under the same controls as the source.

Browser-local processing does not inspect the selected pages for personal data or remove metadata and hidden content. Use a dedicated redaction and document-inspection workflow when disclosure risk matters.

Worked Examples:

References:

• Extract pages from PDFs, Adobe Acrobat, September 23, 2025.
• Allow page extraction from agreement PDFs, Adobe Acrobat Sign, February 2, 2024.