PDF to Excel: extract tables to XLSX online free
Free in-browser PDF to Excel converter. Pulls text out of every PDF page, groups it into rows by y-position and into columns by x-position, then writes a fresh .xlsx with one sheet per page. Powered by PDF.js + SheetJS — no upload.
Drop a PDF to convert
Or click to choose. One file at a time. Stays on your device.
Choose PDFWhen to convert PDF to Excel
Bank statements
Pull transaction tables out of monthly bank PDFs into a spreadsheet for budgeting.
Invoices & receipts
Extract line items into accounting software without retyping.
Reports & data tables
Get research data out of PDF reports for re-analysis in Excel or Python.
Inventory lists
Reformat printed inventory sheets into editable spreadsheets.
How to convert PDF to Excel
Drop the PDF
Drag a PDF onto the page, click to browse, or paste from clipboard. PDF.js parses the document locally — nothing is uploaded.
Pick a page range (optional)
Convert all pages or just 1-5, 3, 1-3,7,10-12. Useful for grabbing the data tables out of a long PDF report.
Click Convert
Text is read from each page, grouped into rows by Y-position and split into columns by X-position clustering. The result becomes one sheet per page.
Download the .xlsx
SheetJS writes a fresh XLSX file ready to open in Excel, Numbers or Google Sheets. Cleanup any merged cells or odd column splits as needed.
PDF to Excel — frequently asked questions
Drop the PDF in, optionally pick a page range, then click Convert. Each page's text is read with PDF.js, grouped into rows by y-position and split into columns by x-position. The result is written to a fresh .xlsx with one sheet per page.
Very accurate for PDFs that contain real text (most reports, bank statements, exports). PDFs that are just scanned images need to be OCR'd first — use our PDF OCR tool. Complex multi-table layouts or merged cells may need cleanup in Excel after import.
No. PDF.js parses the file and SheetJS writes the XLSX entirely inside your browser. The PDF never leaves your device.
Scanned PDFs are images of text with no real text data. Run them through our PDF OCR tool first to add a selectable text layer, then convert with this tool.
How table detection works
A PDF doesn't have rows and columns — it has text fragments positioned at exact (x, y) coordinates. To rebuild a table we use two heuristics:
- Rows. Text fragments whose y-coordinates are within the row tolerance (3 pt by default) are treated as one row.
- Columns. Within each page we cluster every fragment's x-coordinate; clusters within the column tolerance (6 pt) become column boundaries. Each row's fragments are then assigned to the column whose anchor is closest.
When to tweak the tolerances
- Dense tables with closely-spaced rows → lower row tolerance.
- Tables with wide columns and varying alignment → raise column tolerance.
Privacy
Everything runs in this tab. No bytes are uploaded.