PDF Table Extractor
Analyzes the position of text on each PDF page to reconstruct the rows and columns of any table it finds, then lets you download each table as a ready-to-edit Excel or CSV file. It works from the PDF's actual text layer (not OCR), so it handles regular, simple tables better than complex ones with merged cells.
Your file is processed on your device and is never uploaded to our servers.
This tool infers table positions from a PDF's existing text layer (not OCR), based on where text sits on the page — it works well for simple, regular tables, and may need manual cleanup for complex ones (merged cells, multi-line text). Column order always runs left-to-right by position, even for a right-to-left table.
Drag a PDF file here or choose one
One file at a time
What this tool does
Extracts tables from a PDF's text layer. Text items are clustered into rows by vertical position, then split into columns wherever the horizontal gap is unusually large — so the table is reconstructed from where text sits, not read from a ready-made structure.
When to use it
- You need figures from a report's table to analyse in a spreadsheet.
- You want to copy a table without retyping it cell by cell.
- You are moving data out of PDF statements or invoices into something you can process.
How to use it
- Put a PDF on the drop area or choose it.
- Wait for extraction to finish, then review the tables found.
- Export the result for use in a spreadsheet.
A practical example
Input
A financial report containing a simple, regularly spaced 5-column table.
Output
The same table's rows and columns, with columns ordered left to right by their horizontal position on the page.
Limitations
- A PDF has no concept of a table at all; what looks like one is text positioned to suggest it, and this tool reconstructs it by inference.
- Results are good on simple, regularly spaced tables and degrade on merged cells, multi-line cell text or unusual spacing — where columns can be over- or under-split.
- Check the result before relying on it for any calculation; this is not a guaranteed-correct extraction.
- It does not work on tables in scanned documents, which have no text layer.
Privacy and processing
Extraction happens inside your browser; neither the table data nor the file is uploaded.
Technical notes
- Column splitting compares the horizontal gap between consecutive items against that row's own average character spacing, rather than using a fixed threshold that would not suit every font size.
- Columns are always ordered left to right by x-coordinate.
Frequently Asked Questions
No — it relies on the PDF's actual text layer. If your file is a scanned image with no selectable text, run it through an OCR tool first to convert it to text before extracting the table.