How to get a table out of a PDF or a screenshot into Excel
Drop the PDF or picture on the page and press Find the table. Check the column lines drawn over the page, move or add one if a column is wrong, then download Excel or CSV, or copy the table to paste.
Open the file
A PDF of up to 500 pages, a screenshot (paste it with Ctrl+V) or a photo. Choose several screenshots at once to make one table of them.
Find the table
Pages with their own text are read straight from the PDF. Scans and pictures are read with OCR first, in the language you pick. The table is found on each page, and its columns from the gaps between them, or from the lines of a ruled table.
Check the columns
The first page's columns are used on every page. Drag a line to move it, click inside the table to add one, or remove one with its cross. Draw the table again if it found the wrong part of the page.
Save
Set each column to text, number, date or skip, fix any cell in the preview, then download Excel or CSV, or copy the table and paste it into a spreadsheet.
What it fixes for you
- A description that wraps onto a second line stays in its row
- The header repeated at the top of every page appears once
- Page totals and carried-forward lines are left out, unless you keep them
- Ruled tables keep a cell with two lines of text as one cell
Numbers and dates Excel understands
A number column is saved as numbers, a date column as dates, so sums, filters and sorting work straight away. The decimal mark and the order of day and month are found from the whole column, and you can set them yourself.
| In the document | In Excel |
|---|---|
| 1,234.56 · 1.234,56 · 1 234,56 | 1234.56 |
| (12.00) · 12.00- · −12.00 · 12.00 DR | −12 |
| 12.00 CR · $12.00 · 12,00 € | 12 |
| 03/14/2026 · 14.03.2026 · 14 Mar 2026 | 14 March 2026, as a date |
Where does the file go?
Nowhere. Statements and invoices are private, so the reading happens in your browser. The text reader and the table finder download from this site once and are kept in your browser for next time.
Specs and browser support
| Input | PDF up to 500 pages, with text or scanned; JPG, PNG, WebP, GIF (first frame), BMP, AVIF, several at once |
|---|---|
| Output | Excel (.xlsx, numbers and dates typed, header frozen), CSV (UTF-8), or tab-separated text to paste |
| Engine | pdf-lib (MIT) reads a PDF's own text; Tesseract 5 (Apache-2.0) and PaddleOCR PP-OCRv5 (Apache-2.0) read scans and pictures; PP-DocLayout-S (Apache-2.0) on onnxruntime-web (MIT) finds tables; pdf.js (Apache-2.0) draws the pages |
| Download | Table finder 4.9 MB with its 14 to 17 MB runtime; for scans, Tesseract 3.5 MB plus 0.3 to 3.8 MB per language, or PaddleOCR 28 to 39 MB; once, then kept on your device |
| Browsers | Current Chrome, Edge, Firefox and Safari on a computer; phones work for a page or two |
| Price | Free, no account, no limits |
Questions
No. The PDF or picture is read in your browser and stays on your device. Only the text reader and the table finder download from this site, once.
Yes. Number columns are saved as numbers and date columns as dates, so you can add them up and sort by them. 1,234.56 and 1.234,56, (12.00), 12.00- and DR come out as the right number.
Yes. Pages without text of their own are read with OCR first: Tesseract for scans, PaddleOCR for screenshots and photos. Clean scans work best.
Drag the column line on the page, click to add one, or remove one. The preview updates as you go, and every cell can be edited before saving.