Utilore

PDF to Excel

A statement, an invoice, a price list — the numbers are right there, and retyping them is an evening's work. Here the tables are found, shown to you, and handed over as .xlsx or .csv.

A PDF does not store tables — it stores letters with coordinates. The grid below is worked out from where the words sit, so look it over before you rely on the numbers: merged cells come out as ordinary ones, and columns that stand very close together can end up as one.

or drop it here · A PDF up to 100 MB

    The document is read in this browser: nothing is uploaded.

    Also on PDFs: PDF to Word · Image to Text · PDF Editor · Compress PDF

    How to use

    1. Choose a PDF — it is read in your browser, not uploaded anywhere.
    2. Look over the tables that were found: the grid is worked out from the page, not stored in it.
    3. Download the workbook as .xlsx, or take the data as .csv with your separator.

    Good to know

    A PDF does not contain tables

    It contains letters with coordinates. Lines around cells are optional decoration that many generators never draw, and no PDF stores where one column ends and the next begins. So the grid here is deduced: words on the same baseline make a row, and vertical strips that no word crosses make the column boundaries. That works well on real documents and it is still a guess — which is why the result is shown on the page before anything is downloaded.

    What comes out wrong, and it is worth knowing in advance

    Merged cells become ordinary ones — the text lands in the first of them, and the rest stay empty. Columns standing very close together can merge into one, because at that distance the gap between them is no wider than a space. A short table where the same spot is blank in every row can gain a column that was not there. In every one of these cases the numbers stay whole; what changes is which cell they sit in, and that is visible in the preview.

    Numbers stay numbers, everything else stays text

    A cell becomes a number only if it looks like one from end to end: digits, one decimal separator, spaces between thousands. Dates, article codes, phone numbers and ranges go in as text — Excel would happily turn "03.02" into the third of February and "1-100" into a formula, and that is a loss you notice weeks later. Each table gets its own sheet, so a statement with three tables opens as three sheets.

    Frequently asked questions

    Is my document uploaded?

    No. The PDF is read in your browser, and the workbook is built there too. A bank statement or a contract can go through this page.

    It says no tables were found — why?

    Two possible reasons, and the page names the one that applies. Either there is no text at all (the pages are pictures — a scan or a photo), or the text is there but nothing in it looks like a table. Scans can be read with our text recognition tool first.

    Can it handle a table split across several pages?

    Each page is read separately, so a long table arrives as one sheet per page. Joining them in Excel is a copy and paste; deciding automatically where a heading repeats and where a new table starts is exactly where such a guess goes wrong.

    Why .xlsx and .csv both?

    The workbook keeps sheets, numbers and text apart. CSV is one flat file, useful for importing somewhere else — and there the separator matters: Excel in Russian, German or Spanish expects a semicolon, because the comma is busy holding decimals.

    How many pages can it read?

    Up to two hundred in one go. Beyond that a browser spends more time than anyone will wait, and the page says plainly that it read the first two hundred.

    Related tools