How to Turn PDF Tables into Markdown Tables on iPhone (Pro)

The problem

A financial report, a price list or a research paper has the data you need in a table. Copying from a PDF turns it into a column of numbers followed by a column of labels, or one long line per row with no separators. Rebuilding a 40-row table by hand is the kind of job that never ends.

How it works

PDFExtract finds table regions on the page and then works out the grid:

• On a born-digital page, it uses the positions of the words themselves.
• On a scan or image, it runs a layout model to find the table and a table-structure model for the rows and columns, then fills the cells with the OCR text.

All of this happens on the device, using models bundled in the app. The result is a Markdown table you can paste into Obsidian, Notion, GitHub or a note.

Turn it on

1. Unlock Pro.
2. In Settings, the output section’s Rebuild Tables: on.
3. Import (or re-run extraction on an existing project).

Without Pro the switch is visible but locked, and the table’s cells appear as plain lines.

Make it work better

• Use a higher render resolution for small table text.
• Prefer the original digital PDF over a scan of a printout.
• Check header rows and merged cells; Markdown tables can’t merge cells, so they are flattened.
• For wide tables, view the source and tidy it in your editor.

Limits

Detection is model-based, so unusual layouts can be missed or split. Always compare totals against the original.

FAQ

Is table recognition free? No. It is a Pro feature. Free users see the cell text as run-together lines.

Does it work on scanned tables? Yes, using an on-device table-structure model plus OCR for the cell text. Digital tables are usually more accurate.

Where is the switch? In Settings, the output section: Rebuild Tables. It is on by default but only takes effect with Pro.

Will every table come out perfect? No. Merged cells, tiny text and tables without clear lines may need manual cleanup.