WoluTools
Find a tool

PDF & Documents

Convert a PDF to Markdown with headings and page markers

Upload a PDF with selectable text and get Markdown with page markers, plus a page map as CSV. Scanned PDFs without a text layer do not work.

or drop it here

PDF with selectable text · up to 50 pages · 64,000 extracted characters

No file at hand? See a sample result

2 credits per job (about €0.70 with the 20-credit pack) · free account required

You see the cost before the job starts. If the job fails, the credits are released. Credit packs need no plan and do not expire. Prices and credit packs

What you get · fictional sample dataHandbook, page 2
Text in the PDF
Page 2
Returns policy
Send unopened items within 30 days.
Support: help@example.test
Markdown you download
<!-- page 2 -->
## Returns policy

Send unopened items within 30 days.
Support: help@example.test

Also in the download

  • Page-map CSV with the page and position of each block
  • Review workbook listing blocks flagged for checking
  • Evidence page showing each block next to its place in the PDF
BringPDF
GetStructured Markdown, page-map CSV, review workbook and evidence HTML
PrivacyEncrypted source · 24-hour result

One clear job, from source to download

  1. 1

    Add the source

    Supported formats and limits are visible before the upload.

  2. 2

    Confirm the settings

    Review the exact source, options, units and access before processing.

  3. 3

    Inspect and download

    Check the preview and warnings, then unlock the complete package.

Turning a text PDF into Markdown you can check

What PDFs work

You upload a PDF with selectable text, up to 50 pages and 64,000 extracted characters. Scanned PDFs without a text layer are rejected. Run OCR elsewhere first and upload the searchable file. A longer document stops at the limit rather than being cut short without notice, so split it and run the parts separately. This is an AI tool that needs an account and credits, with the cost shown before you start.

What you get and how to check it

You get Markdown with headings and page markers, a page-map CSV, a review workbook and an evidence page. Each block carries its page number and position on the page. The evidence page shows each extracted block next to where it sat in the PDF. Numbers and links are carried through unchanged. The text is extracted first, and the model only labels those blocks. It does not rewrite or summarise.

Tables and uncertain parts

You choose whether tables come out as Markdown tables or HTML tables. Wide tables or tables with merged cells are marked for review instead of being forced into a tidy grid. Where the reading order is unclear, that part is also flagged for review. Go through the flagged blocks in the workbook and compare them with the PDF before you rely on the Markdown.

Questions before you run it

Does it work on scanned PDFs?

No. A scanned page carries no text layer, so the job is rejected instead of being pushed through OCR. Run OCR elsewhere and upload the searchable PDF.

What happens to tables in the PDF?

You pick Markdown tables or HTML tables before the run. Wide or merged-cell tables stay marked as review-required rather than being reshaped into a grid that looks tidier than the source.

How do I check an extracted paragraph against the original page?

Each block carries a page number and coordinates, and the page-map CSV lists them next to the Markdown. The evidence HTML puts the extracted block beside its source location.

Does it rewrite or summarise the text?

No. Positioned text is extracted first and the model only labels blocks that were already pulled out of the file. Numbers and links are carried through unchanged.

What if the PDF is longer than the stated limit?

The job stops at 50 pages and 64,000 extracted characters instead of truncating quietly. Split the document and run the parts separately.