Convert selected PDF pages in your browser, then compare the Markdown with the original. Use this guide when you need to check reading order, recognize an English scan or choose what to keep.
2, 6-8 to convert fewer pages, or choose All pages to reset it. Page numbers count from the start of the file.Follow each column from top to bottom in the PDF, then check whether the draft mixes lines from different columns. Also check heading levels, footnotes and labels. A bold line may still come out as ordinary text. Use Custom layout for separate text columns.
In Page details and conversion options, choose the table draft if one is available. Check that values land in the right cells. The first row becomes the Markdown header and cell line breaks use <br>. Merged cells and automatic joining across pages are unsupported.
Markdown does not reproduce images or interactive PDF features. Check text near diagrams, forms and annotations against the source if it matters. PDF links controls destinations added by the PDF.
English recognition loads when an image page needs it. It works best on clear printed text. Use Compare with PDF to check names and numbers, especially on skewed or low-resolution scans.
Pages with readable text use that text by default. You can try Recognize scanned text (English) in the page’s options, then select Use recognized text for this page if that draft is available. Recognition returns text without reconstructing the original layout, pictures, equations or table relationships.
Open Page details and conversion options, then find the page under Content in this output. You can remove that page or open Keep in Markdown for the options available on it.
Use Remove this page from the output to leave it out of the current result. Change Select pages before converting when you also want to skip processing those pages.
Clear the checkbox for an eligible table to remove it. The option requires a bordered table in native text with boundaries the tool can separate. Surrounding text and captions outside the table stay.
Clear PDF links to remove PDF-added destinations and references. Visible words and printed URLs stay. A Markdown viewer may still make a printed URL clickable.
The result updates when you change these choices. Changing the PDF or page range resets them.
Open Custom layout for the page in Page details and conversion options. Use Text columns for one horizontal region with separate text columns. Use Custom table to mark one simple table region on the page. Coordinates are PDF points measured from the top-left corner. Separate detected bordered tables can remain in the output.
Build the draft and compare it with the source. Review the columns, tables and reading order in the editor before applying it. Changing the region clears that page’s previous layout acceptance.
Use Download Markdown or Copy Markdown once a result is available. The Markdown contains the converted content and any anchors needed for page links. The result names missing pages. The check report lists the included page numbers.
If page 2 fails recognition in a three-page PDF, Download available pages can save pages 1 and 3. Check the page note for the reason, then retry recognition in that page’s options if you need the missing text.
If the tool reports an output check failure, convert again or use Retry output check when it appears. You can save once the current result is available. Files you already saved remain on your device.
Open Page details and conversion options → Download check report for the selected and included pages, your exclusions and any issues the tool found. The checks identify specific risks. Compare the draft with the source to verify the words and their order.
These downloads were captured from the November 7, 2026 converter. Each row pairs the source with the actual Markdown and its separate JSON check report.
| Source and settings | Actual output | What to check |
|---|---|---|
| One-page text PDF Default conversion. |
Markdown · Check report | The leading zeros in 00123 and both example link destinations are present. |
| English scan PDF Automatic English recognition. |
Markdown · Check report | The recognized text contains the invoice amount 123.45 USD. |
| Content selection PDF Page 1. First detected table and PDF links cleared. |
Markdown · Check report | Upper Alpha and the PDF link targets are absent. Lower Alpha and the surrounding text and code are present. |
The November 7 checks included a controlled OCR resource failure and an output check failure. The first preserved the available pages, and the second paused copy and download until the output was rebuilt.
The complete PDF must open before conversion. Choose fewer pages if processing stops early. Page selection cannot bypass a password or repair a damaged file. Your files and results stay in the browser session. Loading the tool or its recognition resources does not upload the PDF.