PDF to Excel

Yes, it uploads - but only for this one conversion. The vast majority of our tools run locally in your browser and the file never leaves your machine; PDF-to-Excel needs server-side parsing of the PDF's internal text and line coordinates to rebuild cells into table structure - doing that across a whole document purely in the browser is slow and unreliable, so this page is one of the upload exceptions. Uploaded files are used only for the conversion: not written to any database, not retained on the server (deleted the moment conversion finishes), never sent to any third party, never seen by anyone. If a file is too sensitive to even pass through an extraction service, do it locally with Excel / WPS: Data - Get Data - From PDF.

No - this page refuses outright and tells you why, rather than handing you an empty sheet. Scans, photos and image-assembled PDFs contain no text layer, only pictures; extracting them requires OCR to recognize pictures into text first, and this tool does not do OCR. The self-test: if you can select and copy the text in a PDF reader, there is a text layer; if selection never latches on, it is a scan. For OCR, run a phone scanning app or OCR-capable office software first.

No formatting is preserved; only content is extracted. A PDF has no concept of a 'table' - only positioned text and lines. This tool rebuilds cells from line and text positions, so fonts, sizes, colors, column widths, merged cells, images, comments, headers and footers are all not reproduced. Tables with visible borders extract most accurately; borderless tables aligned by spacing are attempted, but adjacent narrow columns may merge into one cell - cross-check against the 'full text' worksheet. If you need the layout preserved as-is, use our PDF-to-Word page.

No. Everything in a PDF is text; handed to Excel as-is, Excel would 'guess': long numbers become 1.23E+19 style scientific notation, employee ID 0012 becomes 12 - the leading zeros simply gone. So this tool writes a value as an Excel number only when it is clearly safe (an amount like 1000.00, directly summable once inside); anything with leading zeros, longer than 15 digits, or containing percent/currency symbols is written as text, guaranteed readable as-is. The price is those columns cannot be summed directly - a deliberate trade-off, because a corrupted identifier is far worse than a column you cannot sum.

As many worksheets as tables recognized, named Table1, Table2 in order of appearance; the last sheet is always 'full text', listing the PDF's text line by line so you can check nothing was missed. Across pages: if the next page repeats the header row, the two parts are automatically joined into one table (the duplicate header removed); without a repeated header, a table split across pages becomes two tables and you merge them in Excel yourself - we do not guess here, because a wrong guess is messier than a clean split.

20MB per file - above that the page blocks it before upload (no request is sent, so your data and extraction quota are not spent). 200 pages - the page count is only known after server parsing, so an over-200 file does upload and is then refused server-side, and that attempt does count against your quota. Each IP also has hourly and daily extraction caps; hit them and you are asked to try later. On failure the page names the reason: not a PDF, size limit, too many pages, encrypted (password required), a scan without a text layer, extraction timeout, or service temporarily unavailable. A failed extraction never touches your original file - switch files or split and retry.