Why typed cells matter when you convert a PDF to Excel
Updated 19 August 2026
Converting a PDF table to Excel is only useful if the numbers arrive as numbers. UnboundPDF’s PDF to Excel writes a standard .xlsx workbook in which numbers, dates, percentages and currency are typed cells you can sum and sort, with one sheet per detected table and a Full-text sheet. It runs in your browser; the PDF is not uploaded. It is labelled Lite because it detects tables from text positions, not from ruled lines — yet.
A real .xlsx with numbers, dates and currency as typed cells you can sum — not a wall of text pretending to be a spreadsheet.
UnboundPDF is a free suite of 43 PDF and image tools that run entirely in your browser — merge, split, compress, edit text, OCR in 126 languages, redact, sign, convert and archive to PDF/A. Your document is read and written by the page on your own device; there is no document-upload endpoint in the core tools, no account, no watermark and no daily cap. Every result can be checked — with the Network tab, or with the Document Passport the Workspace writes for a chain of steps.
- Open PDF to Excel and drop in the PDF.
- Export Excel. One sheet per detected table, plus a “Full text” sheet, lands in a .xlsx.
- Open it in your spreadsheet app and check a numeric column: the cells should right-align and sum.
- Need CSV instead? The CSV button is still there.
Text cells vs typed cells
Paste a PDF table into a spreadsheet and you usually get text that looks like numbers. Type =SUM() over it and you get 0, or a column that sorts 1, 10, 2. A typed cell is a number the spreadsheet knows is a number — it sums, averages, formats as currency and sorts correctly. That is the difference this tool is built around: every cell that reads as a number, date, percentage or amount is written as that type.
PDF to Excel is the tool built around this distinction.
What gets typed
Plain numbers, thousands-separated numbers, negatives in parentheses, percentages (12.5% becomes 0.125 with a percent format), ISO and common dates, and currency amounts. European decimals — 1.234,56 — are recognised when the column gives enough signal; a lone “1,5” in a column of plain text is left as text rather than guessed. The rule is: typed when the column agrees, text when it does not, never silently wrong.
One sheet per table, plus everything
Each table the tool finds becomes its own sheet with a header row detected where one exists; a table that continues across pages is joined with the repeated header dropped. A final “Full text” sheet holds the whole document in reading order so nothing outside the tables is lost.
Wrapped cells stay in their table
A cell whose text wraps onto a second line is joined back into its cell instead of being read as a new row or a new table — a small thing that used to break half the tables people actually have.
The CSV button, and why it is safe
CSV is still one click away. Cells that start with the characters spreadsheet apps treat as formulas are neutralised so a malicious PDF cannot plant a formula in your sheet — and real negative numbers and phone numbers are left exactly as they are.
Why it is still labelled Lite
The tool finds tables from where the text sits on the page, not from the ruled lines. Tables with clear columns work well; tables drawn as a grid with merged cells and no whitespace pattern can be missed or mis-split. A ruled-line pass is being built. We do not quote an accuracy figure because we have not measured one we would stand behind. Until both are true, the Lite label stays.
The file you get
A standard .xlsx workbook — Office Open XML, with shared strings and typed cells — validated structurally in our tests. We say “standard .xlsx” rather than naming spreadsheet apps, because the apps are not what we tested; the format is.
Related
Tables to Word instead: PDF to Word. Scanned tables need OCR first. Background: extract text from a PDF and PDF tools for accountants.
Frequently asked questions
Will the numbers be real numbers in Excel, or text?
Typed cells — numbers, dates, percentages and currency are written as values you can sum and sort. A value that cannot be typed with confidence is left as text rather than guessed wrong.
Why is the tool labelled Lite?
Because it detects tables from text positions, not ruled lines. Tables with clear columns convert well; grid tables with merged cells may be missed or mis-split. The label stays until a ruled-line pass ships and accuracy is measured.
Do I get one sheet per table?
Yes, plus a “Full text” sheet with the whole document in reading order. Tables that continue across pages are joined.
Is the PDF uploaded?
No. The table detection and the .xlsx writing happen in your browser; the file never leaves your device.
Can I still get CSV?
Yes. The CSV button remains; formula-like cells are neutralised so a PDF cannot plant a formula, while real negatives and phone numbers are untouched.