Turn a PDF's bookmarks or visible headings into a reviewable outline
Review a report's structure with a separate Markdown list of its PDF bookmarks. Without bookmarks, the tool proposes headings from short text lines whose size and style differ from body text. It does not interpret the document or add bookmarks to the PDF.
Large type might be a chapter title, chart label or decorative quotation. Compare proposed headings with the source. You can correct the downloaded list in a text editor; the on-page result is read-only.
Choose the right starting point for an outline
| Your situation | Recommended approach |
|---|---|
| A PDF already has a bookmark hierarchy | Run the tool to export those titles and their nesting. Existing bookmarks take priority over inferred headings, even if the bookmarks are incomplete. |
| Digital text uses visibly different heading sizes | Try heading detection and compare every result against the source. Short, larger lines are more likely to be treated as headings. |
| Pages are photographs or scans | Local English OCR can obtain text, but it does not recover the original font hierarchy. Expect to build an outline manually unless usable bookmarks already exist. |
| You need the paragraphs beneath each heading | Use PDF to Markdown. This outline export contains heading titles rather than the full document or a summary of its argument. |
Detect two headings in a small digital PDF
The research-notes.pdf sample has two A4 pages with 24 pt headings and 12 pt body text, and no embedded bookmarks. It is designed to make the font-size distinction visible.
- Choose research-notes.pdf from the sample downloads and uncheck Use local English OCR when a scanned page has no readable text.
- Select Create document outline. Read the resulting list and compare it with the headings on both source pages.
- Look for Browser privacy followed by Document checks. Check that the ordinary paragraphs have not become outline entries.
- Download outline and open the .md file in a text editor. Keep the source PDF alongside it if you plan to expand the outline into reading notes.
Expected result. The sample produces a Markdown outline containing the two visible headings in source order. Its body text is omitted from that list. This demonstrates simple typography-based detection, not detection of every possible PDF heading design.
Check the downloaded result
- Compare the count of detected headings with your own review of the document. A small count can be valid for a short file; a large count may include captions or oversized quotations.
- Review indentation. It represents bookmark nesting or inferred relative heading size, not a guarantee that the source author intended those section levels.
- Check for missing sections after scanned pages and read any notice about pages that were not OCR-processed. A total page count does not prove every page contributed headings.
- Check spelling and order, especially where a heading spans several lines. The export does not include bookmark destinations or automatic source-page links.
Troubleshooting
- No outline detected, although the document contains text
The file has no bookmarks and its headings do not differ enough from body text. Long heading lines and list-like text are also excluded by the heuristic.
Try this: Use PDF to Markdown or extract the text and create a short outline manually. Changing a file's visible layout just to trigger detection is unnecessary if you can edit the exported text.
- A scanned document has OCR text but no useful headings
OCR output enters this analysis as plain lines with a uniform style, so the original heading sizes are unavailable.
Try this: Read the recognized text beside the scan and add section labels yourself. Keep OCR as a reading aid; it cannot supply reliable font-based hierarchy here.
- The outline contains old or unexpected titles
Embedded bookmarks are used in preference to detected headings. Their titles can differ from the words currently visible on the pages.
Try this: Compare the PDF viewer's bookmark panel with the result. Correct your downloaded outline or repair the source bookmarks in a PDF editor if the source itself needs updating.
- The PDF will not open or analysis exceeds a limit
Password protection, unsupported structure, file size, page count or extracted-character limits can stop processing.
Try this: Unlock the document with its known password when appropriate, or extract a smaller relevant section. Use the current file and page limits displayed by the picker.
Limits and compatibility
- Structure coverage
- The detector uses bookmark titles and text appearance, not tagged-PDF semantic structure. It bounds the inferred list at 250 headings and collapses immediately repeated heading titles.
- OCR and local handling
- Analysis runs locally. English OCR is optional and limited to 3, 8 or 20 pages by device profile; remaining scan pages can be skipped. The source PDF is not rewritten.
Your files and form values are processed on this device. Downloaded copies remain wherever you save them. See how to check local processing and our privacy policy for site data and optional services.
Common questions
- Will the downloaded outline let me jump to PDF pages?
- No. It is a Markdown list of titles and indentation. Bookmark destinations and page-number navigation are not included in this export.
- Does it summarize what each section says?
- No. A heading list is a structural aid. It does not read charts, explain an argument or decide which sections are most important.
Continue your document workflow
Maintained by Shakti Prasad Ram. Found a mismatch? Report the tool, browser and steps without sending a sensitive document. Back to the tool