PDF & Photo Table → Excel Converter

Turn tables inside a PDF or a photo into a real, editable Excel (.xlsx) spreadsheet, with rows and columns kept intact.

Table in a phone photo or a scan? → Photo or scanned table to Excel (slower, but splits cells better)

Loading the tool...

Why use this

Invoices, transcripts, price lists and government forms are often shared only as a PDF or a photo. Retyping every row into a spreadsheet is slow and error-prone. This tool finds the table, lets you fix anything that looks off right in the browser, and exports a spreadsheet you can open directly in Excel, Google Sheets, or any other spreadsheet app - with no account, no install, and no upload.

When to use it

Move the item table of an invoice PDF into Excel

Choose the PDF and press “Find tables”. Check the result in the preview, then use “Download .xlsx (all tables)”. Values like 1,234 or ₩12,000 land as numbers so you can sum them right away.

Read only the pages that have tables

Enter pages such as 1-3, 5 in “Pages to scan for tables”. This saves time on long files and keeps unrelated tables out.

Turn a photo of a table into a spreadsheet

You can choose an image too. Set “Text recognition (scans)” to “Always”, pick the OCR language and press “Find tables”. Always re-check names and numbers afterwards.

Tidy a table that came out misaligned

Use “Delete column”, “Delete row”, “Merge selected” and “Split focused cell by comma” in the preview, and “Toggle header row” if the first row is a heading.

How to use

  1. Choose a PDF file, or a photo of a table.
  2. Pick which pages to scan and whether to turn on text recognition (OCR).
  3. Review the detected table in the editable preview, then download it as .xlsx or .csv.

How it works

Text and positions are read from the PDF with pdf.js and grouped into rows and columns by our own table-detection code; scanned pages and photos go through Tesseract.js OCR, whose language data is downloaded from this site on first use. The .xlsx file is written directly in your browser (zipped with fflate), so your document is never uploaded. Your file is processed in this browser and is not uploaded. To check this yourself, watch the browser's DevTools Network tab while you use the tool (step-by-step guide). Also available as a free REST API for developers and AI agents (edge-processed, zero retention).

How the table detection works

For text-based PDFs, the tool reads the exact position of every piece of text on the page (via pdf.js), groups text that sits on the same horizontal line into a row, and groups text that lines up vertically across rows into a column. When the PDF also draws visible ruling lines for the table grid, those lines are used to make the column boundaries more accurate. For scanned PDFs and photos, the page or image is rendered and passed through on-device optical character recognition (OCR) to get the position of every recognized word, and the same row/column clustering logic is applied to the word positions.

Editing before you export

The detected table opens in an editable grid. Every cell is a normal text box you can click and retype. You can delete an entire row or column with one click, toggle whether the first row is treated as a header, merge two cells that were split by mistake, and split one cell into several when OCR or text extraction joined two values together. Multiple tables (for example, one per page) appear as separate tabs so you can fix each one before exporting.

Exporting: real .xlsx, CSV, or quick copy

The "Download .xlsx" button builds one real Excel worksheet per detected table inside a single workbook file, using automatic column widths and detecting numeric-looking values (such as 1,234, -12.5, 30% or ₩12,000) so totals and averages work immediately in Excel. If you only need one table, "Download as .csv" saves a UTF-8 file (with the byte-order mark Excel expects, so Korean and other non-Latin text displays correctly) for that table. "Copy this table" copies tab-separated text to your clipboard so you can paste it straight into an existing spreadsheet.

More tips

When you need to convert a PDF table to Excel

  • You received an invoice or a statement as a PDF and need to re-enter the line items into bookkeeping software.
  • You photographed a printed price list or inventory sheet and want it as a spreadsheet you can sort, filter, and total.
  • A government agency or a research paper published statistics only as a PDF table, and you need the numbers for analysis.
  • You have an old scanned document with a table and want to preserve it digitally as a spreadsheet instead of a flat image.
  • A client sent a schedule or a comparison table as a photo in a chat app, and you need it in a spreadsheet to share with a team.

Text PDFs vs. scanned PDFs: why the results differ

A PDF that still has selectable text - for example, one exported directly from Excel, Word, or an accounting system - already stores the exact position of every letter and number on the page. That makes it possible to read the position of each piece of text, line text up into rows by vertical position, and line it up into columns by horizontal position, with high accuracy. If the original document also drew visible ruling lines for the table (the thin lines separating cells), those lines can be used to make the column boundaries even more precise.

A scanned PDF or a photo has no text information at all - it is just a picture of a document. To extract anything from it, the page (or the photo) first goes through on-device optical character recognition (OCR), which tries to recognize each word and remembers roughly where on the page it sits. The recognized words are then grouped into rows and columns the same way as with a text PDF. Because OCR is pattern recognition rather than an exact readout, it is not 100% accurate - a blurry photo or an unusual font can cause it to confuse similar-looking characters, such as the digit 0 with the letter O, or the digit 1 with a lowercase l. Always re-check numbers and names after converting a scan.

Fixing a table that comes out misaligned

Complex tables - especially ones where a single cell spans multiple rows or columns (a "merged cell") - will not always convert perfectly on the first try, because a PDF or an image has no real concept of "this is one combined cell." The editable preview in Dagochim's tool gives you a few simple ways to fix this without starting over:

  • Merge cells: pick two cells that were split by mistake and combine their text into one.
  • Split a cell: if several values got joined into one cell (for example separated by a comma or a tab), split them out into the next cells in that row.
  • Delete a row or column: remove an empty or unwanted row or column with one click.
  • Toggle header row: decide whether the first row should be treated as a header when it is exported.

If a PDF has several pages with tables, each page's table appears as its own tab in the preview, so you can clean each one up separately before exporting everything together.

FAQ

Does it work on PDFs without a visible table?

If the text is laid out in rows and columns it will be detected as a table. Plain paragraph text falls back to one line per spreadsheet row.

Can it read scanned PDFs or photos?

Yes, those are read with on-device text recognition (OCR). Low-quality scans can produce mistakes, so always check the result.

Will complex tables with merged cells convert perfectly?

Simple tables convert well. Tables with multi-line merged cells may need manual touch-ups using the merge/split buttons in the preview.

Is my file uploaded?

No. Table detection, OCR and the Excel file are all handled in your browser. Only the OCR language data is downloaded from this site when a scanned page needs it.

Can I export just one table as CSV or copy it?

Yes. Each table has “Download this table as .csv” and “Copy this table”. “Download .xlsx (all tables)” puts every detected table in one workbook.

Does it work on password-protected PDFs?

A “PDF password” field with an “Open” button appears for those files. Without the password the file cannot be converted.

When should I turn text recognition on?

The default “Auto (scanned pages only)” reads pages whose text is a picture. For photos, or when no table is found, try “Always”; for PDFs that already contain text, “Off” is faster. The language can be Korean + English, English, Korean, Spanish, Portuguese, French or German.

Can I fix the table before I download it?

Yes. You can use "Toggle header row", "Delete column", "Delete row", merge two selected cells, and "Split focused cell by comma". Then click "Download .xlsx (all tables)".

Can I read only the pages that contain tables?

Yes. Enter pages such as 1-3, 5 in "Pages to scan for tables". Leave it empty to read all pages. For a long bank statement, try a few pages first to check the result.

Open-source libraries used

Everything runs inside this browser tab. No file is ever sent to a server. This page uses the following open-source libraries, loaded on demand:

The code that assembles the .xlsx spreadsheet XML itself is written by Dagochim; no separate spreadsheet-building library (such as SheetJS) is used.

Being upfront about the limits

Last checked: October 8, 2026 · Dagochim team

Related tools

Related guide

See the How to Convert a PDF Table to Excel guide for a step-by-step walkthrough.

🗂️ See the full document submission file prep order →