Extract a table from a PDF
A PDF that only "copies" as one garbled line of text, a scanned page saved as PDF, a report with the table buried on page 6 — this reads the visual grid and gives back real rows and columns.
Last updated: September 2026
5 of 5 extractions left this month on the free plan
You can also paste with Ctrl / Cmd + V · JPG, PNG, WEBP up to 10 MB · PDF (Premium) · Select several at once (Premium)
How it works
- 1
Upload the PDF, or a screenshot of it
Premium can upload the PDF directly and pick which page has the table. On Free or Pro, take a screenshot of that page first and upload the image instead.
- 2
The AI reads the grid, not just the text
PDFs often store tables as flattened images or oddly-spaced text runs — a plain "select and copy" collapses the columns into one line. Reading the visual structure avoids that.
- 3
Review and export
Fix any misread cell in the preview, then download a genuine .xlsx workbook or .csv.
Why PDFs are harder than screenshots
A PDF's text layer doesn't always match what you see on the page. Multi-column layouts, merged cells, and scanned pages (a photocopy saved as PDF, with no real text layer at all) all break a simple copy-paste. Vision-based extraction reads the table the way a person would — by its position on the page — rather than relying on the PDF's internal text order.
Multi-page and scanned PDFs
A few things determine how clean the result comes back:
- Premium's page picker lets you choose exactly which page has the table, instead of guessing at page 1
- A scanned PDF (a photocopy or fax saved as PDF) is read the same way as a phone photo — expect to double-check low-confidence cells
- A born-digital PDF (exported from software, with a real text layer) usually comes through cleaner than a scan
- Free and Pro process one page at a time; Premium can upload the full PDF and pick the page
Frequently asked
Upload the PDF directly (Premium) or a screenshot of the relevant page (any plan), and the AI reads the table structure into editable rows and columns — no manual retyping.
Yes — no PDF reader or Adobe product is required. Upload the file or a screenshot of it directly in the browser.
Yes, the same way it handles a phone photo of a printed page. Scans tend to score lower confidence than a born-digital PDF, so review each cell before exporting.
Premium's page picker lets you choose the page to convert, and if that page has more than one table, you pick which one. Add additional pages or tables one at a time and they merge into a single result.
There is no hard page-count limit on the file itself — you pick one page at a time to convert, so a 50-page report works the same as a 2-page one.
Upload the PDF (Premium) or a screenshot of the page (any plan) instead of trying to select and copy the table directly — a plain copy-paste from a PDF usually collapses the columns into one line, while this reads the visual grid and rebuilds real rows and columns.
Related
PDF Table Extractor
For when copy-paste from the PDF falls apart entirely.
PDF to Google Sheets
Send the extracted table straight to a new Google Sheet.
Excel's Get Data From PDF Alternative
For when Excel's own PDF import fails on a scan.
Extract Table from Image
For photos and screenshots rather than PDFs.
Invoice to Excel
For itemized invoices, which are often PDFs.
Pricing
Free tier, plus Pro and Premium for direct PDF upload.