Skip to content
Sırdaş
es · Español
Try it freeToolsAI datasets NewDevelopersServidor MCPCLIPythonn8nTypeScriptAgent SkillSkill: invoicesSkill: contractsBlogBusiness

Extract tables from a PDF to CSV

To pull the tables out of a PDF without uploading it: drop it here and press Extract tables. Columns are rebuilt from the blank space that recurs on every row, which separates right-aligned amounts correctly. You download CSV or Excel, in your browser.

Reviewed on

Sent to servers: 0 BNothing left your device

About this tool

A PDF does not store tables: it stores pieces of text with coordinates. That is why copying and pasting into a spreadsheet usually puts everything in one column.

Here we look for the vertical bands that stay blank on almost every row: that gap is stable even when amounts start at different positions, and it is what really separates the columns. You see a preview of each table before downloading it.

You download CSV, with comma or semicolon delimiter, or Excel (.xlsx) with the amounts as numbers. You can also copy the table as Markdown to paste it into an AI assistant. It needs a PDF with a text layer: a scan without OCR has no coordinates to read.

How it works

  1. 01

    Drop the PDF.

  2. 02

    Choose the format: Excel or CSV (and the delimiter, if CSV).

  3. 03

    Press Extract tables and check the preview.

  4. 04

    Download each table or all of them together.

Questions

How can I check that my file is not uploaded?
Open your browser's developer tools (F12), Network tab, and process a file: no request carries your document. Or load the page, go offline and use it: it keeps working.
CSV or Excel?
Excel (.xlsx) if you will add up or filter: amounts arrive as numbers. CSV if another program will import it or you want plain text.
Excel puts the CSV in one column. What now?
Pick semicolon as the delimiter, which Excel in Spanish locales expects, or download straight to Excel format.
Does it work with a scanned PDF?
No: without a text layer there are no coordinates. Run OCR first (in PDF to Markdown) or use the original PDF the sender produced.
What if the table continues on the next page?
One table is detected per page. Stitching tables across pages is on the roadmap.

More tools