How to convert PDF tables to Excel
- Choose a PDF stored on your device, or load the two-page sample. The sample uses the same extraction and Excel export as your own file. Download it if you want to compare the original document alongside the result.
- Leave Pages empty to process the whole document, up to 30 pages. For a longer PDF, select a smaller range such as 2-8, or separate pages such as 1, 4, 7-9. Duplicate page numbers are processed only once, in document order.
- Choose how Excel should store the values. Keep all values as text is the safest choice for IDs, account codes and mixed regional number formats. Select a numeric format only when it matches the source PDF.
- Select Extract tables. Use the table selector to inspect each result, including every row inside the scrollable preview. Read any page notices below the result. Uncheck Include this table in Excel for a table you do not want.
- Select Save Excel (.xlsx). Open the downloaded workbook in Excel, LibreOffice or another compatible spreadsheet editor. Each included table has its own worksheet; P2_T1 means the first detected table on PDF page 2.
What the converter can recover
PDF pages usually describe where text and lines should be drawn, rather than storing spreadsheet cells. This converter combines those positions into rows and columns. A clear rectangular grid is the strongest input: empty cells remain empty, multiline text stays in one cell, and rectangular merged areas can become genuine merged Excel cells.
Without any ruling lines, the converter looks for at least three aligned text rows separated by consistent empty column gaps. These inferred tables need closer review. Merged cells are not inferred from spacing alone. Separate tables, including tables on later pages, stay separate; repeated headers are kept. Text outside detected tables is not exported.
Two examples to check
The first sample page contains a ruled table with a heading spanning four columns. The heading should occupy the merged range A1:D1. The identifier 00123 stays text with its leading zeros, café and Київ remain Unicode text, and =1+1 remains the literal text =1+1 rather than an Excel formula.
The second sample page contains a simple three-column table without borders. Its Product, Quantity and Price columns should remain separate. With the dot-decimal option, a price such as 12.50 becomes the numeric value 12.5 with two decimal places displayed. With the text option, it stays the text 12.50. Both choices produce real, editable cells.
Numbers, identifiers and formulas
The number options distinguish 1,234.56 from 1.234,56. They apply to the entire extraction, so use text mode when formats are mixed or ambiguous. Excel may display numeric separators according to its own language settings. Dates, currency symbols, percentages and scientific notation remain text; they are not guessed or recalculated.
Leading-zero identifiers, values longer than 15 significant digits and decimals with more than 15 places remain text to avoid Excel rounding. Strings beginning with =, + or @ are never turned into formulas. A simple negative number can be numeric; a formula-like expression such as -1+2 stays text. The tool does not create macros, external links or calculated totals.
Unsupported files and practical limits
Scanned images need OCR, which this tool does not include. A scan with an existing selectable text layer may work if its text positions are accurate. No-text pages, unsupported tables and pages with rotated or vertical text are reported instead of being presented as successful conversions. For mixed documents, only detected tables on supported pages can be saved.
Files are limited to 25 MB and 30 selected pages. Output is limited to 12,000 cells, 50 columns per table, 100 tables and 500,000 characters in total. Excel limits each cell to 32,767 characters. Additional text and drawing limits keep complex PDFs from exhausting browser resources. Try fewer pages when a limit is reached.
Password-protected documents, copying restrictions, damaged PDFs and unusual font encodings can prevent extraction. Export a fresh, selectable-text PDF from the original application when possible. Charts, pictures, colors, fonts and exact page layout are not copied to the workbook. Check important figures and missing cells against the source before using the data.
Frequently asked questions
Is my PDF uploaded?
No. PDF parsing and Excel creation run in your browser. The file is not sent to a conversion service, and the tool does not store it after you leave. Download the workbook before closing the page.
Can I cancel or replace a file?
Yes. Cancel stops the current extraction or export. Remove clears the file and result while keeping your options. Clear also resets the options. Changing a file, page range or value format invalidates the old result, so extract again before saving.
Why are some tables or cells missing?
The detector needs a clear grid or consistent text spacing. Unusual borders, overlapping text, multiple text directions or complex merged layouts may not be recognized. Horizontal-rule tables can retain grouped or multiline headings and empty cells. Superscripts and subscripts are saved as explicit text such as 10^(20) and log_(k), preserving their meaning without calculating formulas. Review every preview and page notice; successful file creation does not guarantee complete extraction.
Will a table spanning two PDF pages be joined?
No. Each detected table is saved separately, with its page and table number in the worksheet name. You can combine compatible tables in your spreadsheet editor after checking repeated headers and column order.
Can I edit the preview?
The preview is for checking and selecting tables. Edit values, fix boundaries and adjust formatting in the downloaded workbook. The output contains real cells, not pasted table pictures or a renamed CSV file.