Split PDF by Table of Contents: Dissertation Chapter Export

What "split pdf by table of contents" really does

The phrase "split pdf by table of contents" describes a specific primitive: take a PDF that has an embedded outline — the tree of headings that shows up in a viewer's sidebar bookmark panel — and use that tree to slice the document into separate files at each outline entry's page boundary. A dissertation with eight chapters and an outline that names each chapter becomes eight chapter PDFs. A 600-page technical manual with sectioned headings becomes one PDF per section. A court bundle whose index lives in the outline becomes one PDF per indexed item.

The reason this is a useful primitive — rather than just "split at page N" — is that the author of the PDF already did the work of identifying the chapter boundaries. They live in the outline as machine-readable metadata. A table-of-contents-driven splitter reads that metadata and produces correctly-sized output files without anyone having to scroll through the source document noting page numbers.

PDF outlines: the underlying data structure

A PDF outline is a hierarchical tree of outline items stored in the document catalog. Each outline item has a title (the human-readable heading), a destination (the page and view it points to), and optional children (sub-headings). The structure is defined in section 12.3.3 of the PDF specification ISO 32000-1, and the open-source pdfcpu reference at pdfcpu.io documents the outline-manipulation commands in detail. The pdf-lib library at pdf-lib.js.org exposes the same tree at the JavaScript level, which is what makes browser-local outline splitting feasible.

For PDFs born from LaTeX (via hyperref), Microsoft Word (via Save As PDF with heading bookmarks enabled), Pages, InDesign, or LibreOffice, the outline is populated automatically from heading styles in the source document. For PDFs born from scans, image-to-PDF conversions, or simple print-to-PDF without heading metadata, the outline is empty and needs to be added manually before outline-splitting can work.

The outline tree is what your reader sees when they expand the bookmarks panel. Open any dissertation or technical manual in a PDF viewer, expand the panel, and the tree of chapter and section headings is the data the split-by-outline tool reads.

Dissertation workflow: chapter-by-chapter export

Most graduate schools and university libraries accept dissertation submissions as a single combined PDF and as separate-chapter PDFs. The combined version is the archival copy; the separate chapters are for committee review, partial publication, and ProQuest indexing. The split-by-outline workflow exists to bridge those two formats without manual chapter extraction.

The typical setup. A doctoral candidate writes the dissertation in LaTeX or Word with proper chapter headings. They export to PDF with bookmarks enabled — thehyperref package in LaTeX does this by default with the bookmarksopen=true option, and Word's Save As PDF dialog has a "Create bookmarks using headings" checkbox under Options. The output PDF has an outline tree that mirrors the chapter structure.

Then committee review starts. Each committee member wants only the chapters in their area of expertise. The methodology expert wants Chapter 3 (Methods) and Chapter 4 (Data). The theory advisor wants Chapter 1 (Introduction) and Chapter 2 (Literature Review). The candidate could email the full 300-page PDF every time, but it's easier — and far more storage-efficient on the reviewer's end — to send only the relevant chapters as standalone files.

Running split-by-outline at level 1 produces one PDF per chapter, named "Chapter-1-Introduction.pdf", "Chapter-2-Literature-Review.pdf", and so on. The candidate selects the relevant files and attaches them to the committee email. The whole workflow — from running the splitter to having chapter-named files ready to attach — takes under a minute on a typical laptop.

Step-by-step walkthrough

1. Confirm the source PDF has an outline. Open the dissertation in your PDF viewer and expand the bookmarks panel. You should see a tree with chapter and section headings. If the panel is empty, the outline is missing — jump to the "When the PDF has no outline" section below.
2. Open the split-by-outline tool. Navigate to pdfmavericks.com/split-by-outline. The tool loads in your browser — no account, no upload.
3. Drop the dissertation PDF. Drag the file into the upload zone. The tool reads the outline via PDF.js and renders a preview of the outline tree on the right side of the screen.
4. Pick the outline depth. For dissertations, depth 1 is almost always correct — one output PDF per chapter. For technical manuals with deep nesting, depth 2 produces one PDF per section. The preview updates to show how many output files each depth setting would produce.
5. Run the split. Click the "Split by Outline" button. The tool walks the outline tree, identifies the page range for each entry, and uses pdf-lib to write each range as a separate PDF.
6. Download the results. The tool produces a ZIP archive containing all the chapter PDFs, named after the corresponding outline entries. The Save dialog appears with a default filename like dissertation_chapters.zip. Save it, extract, and the chapter-by-chapter files are ready to send.

For a 300-page dissertation with 8 chapters, the entire operation completes in 2 to 4 seconds in a modern browser. The bottleneck is the ZIP archive creation, not the PDF splitting itself — page references are copied structurally, not re-rendered.

Choosing the right outline depth

The depth setting controls how deep into the outline tree the splitter goes before it cuts. Depth 1 means "treat each top-level entry as a split boundary". Depth 2 means "treat each level-2 entry as a boundary, ignoring the level-1 grouping". Depth 3 means level-3 entries. And so on.

For dissertations, depth 1 is the default. The level-1 entries are the chapters, and chapter-sized output files are what graduate-school and committee workflows expect. Splitting at depth 2 produces section-sized files (Chapter 3.1, 3.2, 3.3, ...) which is almost always too granular for review purposes.

For technical manuals — software documentation, API references, engineering handbooks — depth 2 is often the right setting. The level-1 entries are part-sized groupings ("Part I: Getting Started", "Part II: Advanced Topics") which are too coarse, and level-2 entries are individual chapters or major sections, which match the unit of distribution. Splitting an Apache HTTP Server documentation PDF at depth 2 produces one PDF per major topic.

For court bundles and discovery productions, the right depth depends on how the bundle was indexed. A bundle indexed as "Witness Statements / Witness A / Statement 1" would split at depth 3 to get individual statement PDFs. A bundle indexed as "Document 1", "Document 2", ... at the top level would split at depth 1.

When the PDF has no outline

If the bookmarks panel is empty when you open the PDF, the outline tree is missing and outline-splitting will return a single file (the original) or fail with a message saying no outline was found. Three recovery paths.

Re-export from the source. If you have the original Word, LaTeX, or InDesign file, re-export with bookmarks enabled. In Word: File > Save As > PDF > Options > check "Create bookmarks using Headings". In LaTeX with hyperref: add \usepackage[bookmarks=true]{hyperref}and rebuild. The resulting PDF will have an outline that matches the heading structure in the source.

Add the outline manually. Open the PDF in Adobe Acrobat Pro, PDF-XChange Editor, or a similar tool. Navigate to each chapter start, right-click in the bookmarks panel, and add a bookmark entry. This is slow but it's a one-time cost — once the outline is in place, splitting becomes a one-click operation forever after.

Script the outline addition. For documents with a predictable heading pattern, pdfcpu can generate an outline from a JSON file describing the chapter boundaries. The pdfcpu bookmark command reads the JSON, embeds the outline, and writes a new PDF with the bookmark tree in place.

If none of these is practical, fall back to the standard split tool with explicit page ranges or split-every-N-pages mode. The output won't be chapter-named, but the page boundaries will be correct.

Manuals, court bundles, and other use cases

Software documentation. A 1,200-page Spring Framework reference PDF has a deep outline. Splitting at depth 2 produces one file per major topic (Beans, AOP, Transactions, Web MVC, Security), each in the 50 to 150 page range. Reviewers and trainees get only the parts they need rather than the full reference.

Equipment manuals. Maintenance manuals for industrial equipment often run 500+ pages with sections for installation, operation, maintenance, troubleshooting, and parts catalog. Splitting at depth 1 produces one file per section, which is how field technicians actually use the documentation — grab-and-go reference, not a single megafile.

Court bundles and discovery productions. A typical commercial litigation bundle runs 500 to 5,000 pages with an outline indexing every document. Splitting by outline produces one PDF per indexed document, which is the format counsel actually works with during witness prep and cross-examination. Combined with bates numbering, this is the core legal-tech workflow for bundle prep.

Regulatory filings. Indian MCA filings, SEBI prospectuses, and RBI master directions are often distributed as single multi-hundred-page PDFs with section outlines. Compliance teams split them by outline to assign sections to different reviewers in parallel.

Why the split runs in your browser

The documents people split by outline are often the most sensitive PDFs in their workflow. Unpublished dissertations, embargoed research, draft court bundles, IP- protected manuals, pre-filing regulatory drafts — uploading any of these to a third-party splitter creates a copy on a server with retention policies you don't control. The pdfmavericks.com split-by-outline tool runs entirely in your browser tab using PDF.js and pdf-lib. The PDF is read from local disk via the File API, the outline is parsed in memory, the chapter PDFs are written in memory, and the result is delivered to disk through the standard Save dialog.

No network request carries PDF bytes. You can verify this in the browser's Network tab — open developer tools (F12), switch to Network, check "Preserve log", run the split, and confirm there is no multipart POST with file data. The only network traffic is the initial page load: HTML, JavaScript chunks, fonts, and a small WebAssembly bundle for compression. For more on the architecture, see the no-upload PDF tool overview.

For dissertations and research bundles in particular, the browser-local approach also avoids the question of cross-border data transfer. A doctoral candidate running the splitter in India is not transmitting their unpublished work to a US or EU server, which simplifies the answer to any future Subject Access Request or DPDP Act compliance review.

For other PDF operations on the same source — combining specific chapters back together, reordering pages, removing draft sections — the organize-pdf workflow covers the structural-editing primitives without leaving the no-upload model.