PDF bookmarks viewer — how the outline tree works and what to do with it
Most long PDFs (books, technical manuals, regulatory filings, theses) carry a bookmark tree — a hierarchical outline that lets readers jump to chapters, sections, and sub-sections from the side panel of their PDF reader. Bookmarks are easy to overlook, but for a long document they are the single biggest factor in whether the file is navigable or not. Here is how the bookmark structure works, why some PDFs have it and others do not, and the three places where surfacing the bookmark tree as standalone data is genuinely useful.
What a bookmark tree is
Every PDF can include an /Outlines dictionary — a tree structure where each node has a title, a destination (which page to jump to, where on the page), and zero or more children. PDF readers display this tree in their side panel: Acrobat calls it Bookmarks, Preview calls it Table of Contents, Chrome calls it Document outline. Clicking any entry jumps to the destination.
The tree is independent of the document's visible content — you can have headings on every page without bookmarks, or bookmarks pointing to specific pages without visible chapter titles. Most well-authored documents have both: visible headings + matching bookmarks. Many auto-generated PDFs have only one or the other.
Why some PDFs have bookmarks and others don't
The bookmark tree is opt-in at PDF generation time. The source application has to explicitly create the /Outlines dictionary. Different generation pipelines have different defaults:
- Microsoft Word "Save as PDF". Auto-generates bookmarks from Word's heading styles (Heading 1, Heading 2, etc.). If your Word doc uses styled headings, the PDF has bookmarks. If you used bold text without styles, no bookmarks.
- LaTeX with hyperref package. Auto-generates bookmarks from \section, \subsection, \subsubsection structure. Standard for academic papers and technical books.
- Scanned PDFs. Almost never have bookmarks — scans are usually image-only with no structural awareness. Add bookmarks manually with Acrobat Pro, or use AI · Mindmap to infer a hierarchy and convert it to a bookmark structure.
- Browser "Print to PDF". Usually no bookmarks. Browsers focus on rendering the page; they do not extract heading structure into the PDF's outline.
- Hand-edited PDFs. Whatever the original generator produced, minus anything an editor removed. Always worth verifying with this tool after a long edit.
What our viewer surfaces
The PDF Bookmarks viewer parses the /Outlines dictionary and renders it as an indented tree:
- Title of each bookmark. The text that appears in PDF readers' side panel.
- Depth (indentation). Top-level chapters at depth 0, sections at depth 1, etc. The indentation matches how the reader's bookmark panel would show it.
- Destination page. Which physical page the bookmark jumps to.
- Total bookmark count. A quick number to gauge whether the document has comprehensive bookmarks or just a few top-level entries.
Three real-world use cases
Where surfacing the bookmark tree pays off:
- Audit before redistribution. If you are about to share a long PDF, viewing its bookmark tree confirms it is navigable. A 400-page report without bookmarks is a usability disaster; the viewer surfaces this before recipients hit it.
- Build a navigation index for a website. If you host a PDF for download, copying its bookmark tree into a table-of-contents block on the download page gives visitors a preview of what is inside. The viewer's output makes this a copy-paste step.
- Migrate a long document to web. Converting a 500-page PDF to a series of web pages? Start by exporting the bookmark tree — it is the natural page-segmentation map. Each top-level bookmark becomes one web page; sub-bookmarks become headings within those pages.
Two patterns worth knowing
Friction points worth knowing before the viewer surprises you:
- Bookmark titles don't always match page headings. The bookmark title is whatever was supplied at generation time. Sometimes it matches the page heading verbatim; sometimes the author shortened it for the side panel. Do not assume the two are identical.
- Empty bookmark trees are valid. Some PDFs technically have an /Outlines dictionary but with zero entries. The viewer reports "no bookmarks" rather than an error. If you expected bookmarks and don't see them, the generator probably didn't create them — not a tool failure.
Limits and compatibility
On the free web tool, the bookmarks viewer handles PDFs up to 100 MB with no bookmark-count cap. Parsing runs in your browser; the file never leaves your machine. Output renders the tree visually and exports as JSON / Markdown / CSV for downstream use.
Common pairings: Pair with PDF Inspector for a complete document audit (pages + bookmarks + fonts + size). Pair with Extract Pages to slice the document along the bookmark structure when migrating to web or chunking for search.