PDF Splitter

Introduction:

A PDF often becomes the shared version of a document after the real work is already finished. That does not mean every recipient should receive the same page set. A scanned stack may contain several forms, a board pack may need committee sections, and an upload portal may accept only the pages relevant to one request.

PDF splitting creates smaller PDFs from selected source page positions. It is different from deleting pages in the original file: the source can remain unchanged while each new file gets its own page order, filename, delivery boundary, and review purpose.

The page number used for splitting is the source page position, counted from the first page inside the file. Printed labels can say Cover, i, A-1, or 3, but the range 1-3 still means the first three pages in the PDF. Covers, dividers, Roman-numeral front matter, and scanned page labels are common reasons for visible page numbers to drift from file positions.

Good split plans name the boundary as well as the pages. Legal exhibits, invoice groups, forms, chapters, appendices, and review packets each need a different division. Repeating a cover page can be intentional when every packet needs context, but accidental overlap can create duplicate evidence or confuse a recipient who expects each packet to be unique.

Splitting also has document-feature limits. Bookmarks, page labels, digital signatures, form behavior, attachments, reading order, and accessibility tags can depend on structure outside the copied page artwork. Important packets should be opened after splitting and checked at the first page, every boundary, and the final page.

How to Use This Tool:

Use the splitter when one PDF needs to become several smaller PDFs in the browser. Build the plan first, then check the manifest before downloading the ZIP file.

1. Choose or drop one PDF file. If multiple files are selected, only the first usable PDF is loaded and the extra files are ignored.
2. Select Split mode. Custom range packets creates one output PDF per token, Every page as a PDF creates one file for each source page, and Fixed pages per PDF creates same-size chunks with a shorter final packet when needed.
3. For custom ranges, enter tokens such as 1-3, 4-6, 7-, -4, odd, even, or all. After a PDF is loaded, Halves, Thirds, and Odd/even can fill common range plans.
   If a range exceeds the loaded page count, starts after it ends, or leaves an empty token, fix the range before splitting.
4. Set Output prefix so the generated PDF and ZIP filenames make sense outside the browser. Use Output metadata, Packet guard, and Browser work limit in Advanced when the handoff or file size needs tighter control.
5. Review the summary, warnings, Split Manifest, Split Packet Map, and Split Checks. Confirm page counts, overlaps, packet guard status, and work-limit status before generating files.
   The selected PDF has a 150 MB load cap. The adjustable browser work limit is a planning guard and does not raise that hard file cap.
6. Press Split PDF. When Split Package shows generated PDFs packaged in one ZIP, download the ZIP and keep the manifest, checks, CSV, DOCX, or JSON only when you need an audit record.

Interpreting Results:

The ZIP file is the deliverable, but the manifest is the evidence that explains what the ZIP contains. Treat the result as ready only when the planned page map matches the document boundaries you intended.

• Split Package reports split status, source file, source pages, split mode, planned packets, overlapping pages, ZIP filename, ZIP size, and privacy path.
• Split Manifest lists each packet, range token, source page positions, page count, filename, generated size, and status.
• Split Packet Map charts pages per output PDF so long packets, short tail packets, and uneven splits are easier to spot.
• Split Checks shows whether the PDF runtime, ZIP runtime, source file, work limit, split plan, packet guard, overlap state, chart, and output are ready or need attention.
• JSON records settings, source details, packet rows, warnings, checks, and generated output details for machine-readable review.

An overlap warning means at least one source page appears in more than one output PDF. That can be correct for shared covers or instruction pages, but it should be checked because accidental repeats can confuse filing, review, billing, or evidence records.

A clean plan does not prove the document content is correct. Open the generated PDFs when page boundaries matter, especially around scanned separators, attachments, signatures, forms, nonstandard page labels, and accessibility-sensitive documents.

Technical Details:

PDF splitting starts with the loaded page count. Each split mode expands into packet definitions, and each packet contains an ordered list of source page positions. The generated PDF copies pages from that list, so repeated source pages become repeated placements in separate output files.

Range tokens are validated after the source page count is known. Whitespace around tokens is ignored, tokens are separated by commas, and a token that points past the end of the document is rejected rather than silently shortened.

Fixed Packet Formula:

For fixed-size chunks, the planned packet count is the ceiling of source pages divided by pages per PDF:

For example, a 23-page source split at 5 pages per PDF creates ceil(23 / 5) = 5 packets: four five-page packets and one three-page packet.

Range Token Rules:

Mode Behavior:

Validation Boundaries:

Limitations and Privacy Notes:

The selected PDF, generated PDFs, manifest data, and ZIP bytes are handled in the browser session. The page still needs its PDF-copying and ZIP-creation support to load, but the selected PDF is not uploaded for splitting.

Page copying should not be treated as preservation or certification for every PDF feature. Document-level bookmarks, page labels, attachments, form relationships, signatures, accessibility tags, and reading order may need separate inspection in the generated files.

Large PDFs, damaged PDFs, password-protected PDFs, and heavily interactive PDFs can fail or produce files that need manual review. For legal, archival, accessibility, or compliance work, keep the original file and record the manifest used for each delivered packet.

Worked Examples:

These cases show how page-position ranges become planned packets and review warnings.

Advanced Tips:

• Use Load sample to test custom ranges, overlap warnings, and ZIP generation before handling a real document.
• Lower Packet guard when a document should produce only a few packets; it catches accidental every-page plans before generation.
• Set Output metadata to neutral when the output PDFs should not carry a title based on the source filename.
• Use Split Packet Map to catch uneven fixed chunks, a very large custom packet, or an unexpected one-page tail packet before downloading.
• Save the manifest CSV, DOCX, or JSON when another person needs to confirm which source page positions went into each delivered PDF.

FAQ:

Glossary:

Source page position

The page number counted from the start of the PDF file, independent of printed labels.

Packet

One planned output PDF created from a range token, one source page, or a fixed-size chunk.

Range token

A short page-selection instruction such as 5, 8-12, 7-, odd, even, or all.

Open-ended range

A range where the beginning or end is inferred from the source document, such as 7- or -3.

Overlap

A repeated source page placement that appears in more than one output PDF.

Manifest

The review table connecting packets to source pages, page counts, filenames, generated sizes, and status.

References:

• PDF Standards, PDF Association.
• File API, World Wide Web Consortium, 3 December 2025.
• Using files from web applications, MDN Web Docs, 18 September 2025.
• Blob: arrayBuffer() method, MDN Web Docs, 2 December 2023.