PDFGym
EN

Turn PDF tables into Excel

or drop one here
Choose a PDF first.

Financial tables locked in PDF are a modeling project waiting to happen. This page detects the grids in the native text, packs each selected one into its own worksheet, and reopens the workbook to confirm names, rows, and cells before download. Every value lands as text, with no macros or hidden sheets aboard.

The annual report that had to become a model

When the figures live in a report instead of a spreadsheet, copy-paste is where errors breed. Extract the grids cleanly and model from those.

An analyst opens annual-report-2025.pdf, 40 pages, hunting the balance sheet, income statement, and cash flow tables. The detector lists every grid with its page and shape, all pre-ticked. She unticks two irrelevant schedules, keeps the three statements, and presses Convert to Excel.

The download arrives as annual-report-2025-tables.xlsx: three visible worksheets named for their pages, values as text, proof counting sheets, rows, and matched cells. Her model starts from clean imports instead of a weekend of retyping.

Four passes from PDF to workbook

  1. Open the PDF

    Press Choose a PDF, or drop the file on the page. The tables need selectable text to stand on.

  2. Tick the tables

    The Choose detected tables panel cards every find as Worksheet N with its page, row, and column counts. All start ticked; Select all and Deselect all reshape the set fast.

  3. Build the workbook

    Press Convert to Excel. Each ticked table becomes its own visible worksheet in a single XLSX file, every value stored as text so nothing executes on open.

  4. Check and download

    The workbook is reopened and its sheet names, rows, and cells matched against detection, with macros, formulas, external links, and hidden sheets confirmed absent. Download Excel then saves the verified file.

Inside the workbook you download

Sheets carry honest names: Page 12 Table 1 means the first grid found on page 12, trimmed to the 31 characters Excel allows. Grids are never fused, so each sheet mirrors exactly one detection, and the result pairs the source page with the first worksheet for a visual verdict on the inference.

Cells hold text, full stop. A cell reading =1+1 stays the four characters it was, never a computed two, which is why the file opens without security prompts. The proof counts reopened sheets, rows, and matched text cells, and the result note confirms the absence of formulas, macros, external links, and hidden sheets.

Cards preview each find before you commit: Worksheet N with its page and row-by-column shape, all pre-ticked so a full report converts in one press. Deselecting is instant, and the first worksheet renders beside its source page in the result for a last visual verdict.

Where the sheets need human help

Detection guesses boundaries, and four layouts fool it: merged headings pinned to one column, sideways tables, scattered form fields, and article prose posing as a grid. Scans refuse with the reason named, since pixels give the aligner nothing to group. None of these get an in-tool fix, so correct the stragglers in Excel after import.

Text cells are the other adjustment. Sums and charts need numbers, so convert the columns inside Excel before modeling. That extra step is the price of a file that can never execute anything, and it beats explaining a live formula to an auditor.

When a whole page refuses, look for the scan: a photographed schedule carries no text layer, so the detector reports nothing found rather than an empty sheet. Run that page through OCR elsewhere first, then bring the text version back.

When a workbook is overkill

Pipelines and notebooks usually prefer PDF to CSV, which runs the same detector and adds a formula-guard mode for risky cells. Whole-document prose belongs to PDF to Markdown, page pictures to PDF to images. Going the other way, Excel to PDF turns a worksheet back into a PDF.

Why money workbooks stay on your device

Statements, payroll grids, and client figures should not tour a stranger server on their way to Excel. Detection, writing, and the reopen check all happen where the PDF already sits, with no account and nothing retained.

Text-only cells extend that caution past the download. No formula can fire, no macro can hide, no external link can phone home, because the format forbids their presence and the check confirms it. The workbook opens clean on any machine you trust.

Short answers about PDF to Excel

Why do my numbers arrive as text?
By design. Values are stored as text and never as formulas, so nothing executes when the workbook opens. Convert the columns inside Excel before summing or charting.
What do the sheet names mean?
Page 12 Table 1 is the first grid found on page 12, counting through the document in order. Names trim to the 31 characters Excel allows.
Can I get one big sheet instead of many?
No. Each table becomes its own worksheet so unrelated grids never fuse. Copy the sheets together in Excel when one view serves better.
Will Excel warn about macros?
No. The writer creates no macros, formulas, external links, or hidden sheets, and the check confirms their absence. The file opens without security prompts.
Same tables as the CSV tool?
Yes. Both pages run one detector, so the finds match. Pick CSV for pipelines with its formula-guard mode, Excel for modeling in worksheets.
How many tables fit one workbook?
No fixed cap stands in the way; the practical ceiling is device memory, and heavy reports simply take longer. Deselect the schedules you will never model to keep the file lean.

More tools for number-heavy files

The tools below handle figures the same safe way: converted on your device, rechecked before download, and never asking for an account.