PDF to Excel
Convert PDF tables to an Excel XLSX workbook in your browser. Ruled tables are found page by page with PyMuPDF, and the PDF is not uploaded.
Upload PDF File
Read the full guide: How to Convert PDF to Excel Without Uploading the File
About This Tool
PDF to Excel finds the tables in a PDF and writes them to an .xlsx workbook. Add one PDF and select Convert to Excel: PyMuPDF's table finder checks every page, and pandas with openpyxl builds the workbook, all running as Python inside your browser tab. The workbook is named after your PDF.
Every page with at least one table gets its own sheet, named after the PDF page, such as Page 1 or Page 4. Several tables from one page are stacked on that sheet with an empty row between them, and the first row of each table becomes the header row. Only the cell text is copied; fonts, colours and shading stay behind. If nothing is found anywhere, the workbook holds a single Info sheet saying that no tables were detected.
Detection works from the lines drawn on the page, so tables with visible borders or rules are the ones it finds reliably. Tables laid out with spacing alone can be missed or split, and scanned pages offer neither text nor lines to work from. Numbers usually arrive as text, so convert them in Excel before sorting or adding up. For borderless tables, Extract Tables has its own detection settings and CSV output.
The first run downloads nearly 50 MB of Python engine files from this site, so expect a pause with status lines such as Installing Pandas; the PDF itself stays in the tab and is not uploaded. When the figures will feed a trend analysis, a product quality review or a submission table, check every value against the source PDF before you rely on it.
Step-by-step instructions and common mistakes: How to Convert PDF to Excel Without Uploading the File.
How to Use
Add one PDF
Drag a PDF onto Upload PDF File or click to browse. A note confirms that each page's tables go to separate sheets.
Select Convert to Excel
There are no settings. Status lines such as Loading Python environment appear while the engine loads, then the conversion runs.
Download the workbook
Select Download to save the .xlsx file, which carries the same name as the PDF.
Check the sheets
Open each Page sheet next to the PDF. Fix headers, join tables that ran over a page break and convert text numbers to numbers.
Use Cases
Reusing results tables
Pull ruled results tables out of a supplier certificate of analysis or a published study so the figures can be charted or compared.
Reconciling statements
Bring a bordered statement or invoice table into a spreadsheet to match it against your own records.
Building a comparison sheet
Collect specification tables from several product data sheets, one PDF at a time, into workbooks you can combine.
Frequently Asked Questions
How is the workbook organised?
There is one sheet for each page that contains a table, named after that page, such as Page 3. Pages without tables get no sheet. Several tables on the same page share a sheet, one below another, with a blank row between.
Why did I only get an Info sheet?
No table was detected in the document. The finder looks for drawn grid lines, so tables made from spacing alone, or tables in a scan, are often missed. Try Extract Tables, which lets you adjust the minimum rows and columns.
Can it read a scanned PDF?
Rarely. A scan is a picture with no text layer and no drawn lines. OCR PDF can add recognised text, but it does not add the ruling lines the finder relies on, so the table may still not be found.
Why are my numbers stored as text?
The converter copies the text it reads in each cell, so figures arrive as text rather than numeric values. Use Excel's convert to number option before you sort or calculate.
Why is the first conversion slow?
The converter runs Python in the browser and has to download about 48 MB of files first, including PyMuPDF and pandas. Your browser normally caches them, so later runs start sooner.
Does it work with password-protected PDFs?
Remove the password first with Decrypt PDF, which needs the current password, or the conversion may fail.