SELECT THE TABLE
Compare the original page with your rows
After choosing a PDF, its original page is rendered locally. Draw around a table, including at least three populated rows. Review whether the first row is headings or data after extraction. Zoom or rotate the view without moving your selection. You can also select the whole visible page, use the arrow keys or enter exact coordinates.
Leave space around the text: a boundary that clips a cell blocks that extraction. Clear the region to return to full-page or page-range extraction. On smaller screens the page and table stack vertically.
SUPPORTED INPUT
Which PDF tables can be extracted?
This tool is designed for digital PDFs that contain selectable text arranged in a stable table. Being able to select text is a useful first check, but it does not guarantee support: the page must also fit the engine's bounded text and layout subset.
No silent missing words
Supported text inside the inferred table area or your selected rectangle must be accounted for in the cells. If text was left out, incomplete spreadsheet output is blocked. For descriptions split into several text fragments on the same line, use Adjust columns to set boundaries, then extract again. Wrapped lines and sparse rows can be ambiguous: this tool does not guess which record they belong to. Compare the entire result with the original page, including anything outside the inferred table area.
EXCEL · XLSX · CSV
A spreadsheet you can actually work with
Every selected table becomes a separate Excel sheet, with a frozen header and column filters. Prefer CSV? Download a ZIP with one rectangular CSV per table. Or select a table and copy its rows directly into your spreadsheet.
Keep IDs such as 00123 intact
For example, if a supported source cell contains 00123, the Excel download stores 00123 as text—not the number 123. A value such as 03/04 also stays text instead of guessing a date. These are illustrative values, not a promise that every PDF layout can be read.
For CSV or copied rows, choose Text for ID and date-like columns in your spreadsheet's import settings. CSV alone cannot prevent Excel or Sheets from changing the values. After comparing the result with your PDF, convert only the intended numeric columns if you need sums.
MULTI-PAGE TABLES
Combine matching tables into one Excel sheet
When one table continues across several pages, extract those pages and review their column headings. Choose the tables you need, update the downloads, then expand Combine matching tables into one sheet. Download one combined Excel file or one CSV, with source PDF page and table columns for every row.
Combining is optional: ordinary Excel downloads still keep one sheet per table. The combined export requires exactly matching column names, order and row widths. It writes the headings once and preserves all data rows, including duplicates. A data row that repeats the headings is kept and flagged, not silently deleted. Tables with matching headings can still describe unrelated datasets: decide whether they belong together before combining.
A continuation page has no repeated headings?
Check Check what the first row means in the result. Its original cells are shown before any header renaming. For that continuation table, choose first row as data and use the headings from a reviewed matching table, then apply the choices. The first record returns to the data rows; combined exports can include it. If there are no headings anywhere, choose generated column names and review each column yourself. A new PDF or changed extraction settings clears these choices.
Try Matching multi-page table under the example PDFs, compare both pages, then inspect the combined spreadsheet. This is not OCR, automatic data cleaning or accounting reconciliation.
FREE · PRIVATE · NO SIGNUP
Your PDF is processed locally
The selected PDF bytes are read in this browser. The extraction engine has no upload or remote-fetch path, and you do not need an account. The page may still request its ordinary static application files.
NO OCR
Scanned and image-only PDFs are not supported
This tool does not turn page images into text. Encrypted files are BLOCK. Scanned, rotated, skewed, merged-like, unresolved-font and unsupported stream surfaces return INDETERMINATE instead of invented rows.
METHOD
What PASS means
PASS means the selected pages contain a stable table inside the bounded digital-text operator subset, with no unsupported coverage signal. It is not a complete ISO 32000 validation or an accuracy certificate.
SPREADSHEET SAFETY
Formula-safe spreadsheet exports
CSV neutralizes leading formula characters. XLSX uses inline string cells and also neutralizes formula-leading strings. Raw extracted values remain visible in the JSON analysis for review.
REVIEW BEFORE YOU EXPORT
Check more than the first few rows
Choose an extracted table, search for an ID or any cell value, and page through every row beside the original PDF. Original data-row numbers stay visible when you search. Use First, Previous, Next, Last or a page number to move through large tables.
Search is a review aid, not an export filter: your spreadsheet and copied table still contain all rows in their selected tables. Select only the tables you need before rebuilding downloads. Technical cell coordinates load when you open the details panel, 100 records at a time.