Skip to main content
PPDF123

Read a Scanned PDF Table into Excel

Scanned PDF to Excel renders image-only pages, recognizes their words with Tesseract in the language you choose, rebuilds rows from the word positions and lets you correct them before saving an Excel workbook. Accuracy depends on scan quality.

Drag and drop files here, or click to choose

Accepts PDF · up to 500.0 MB per file · up to 20 files at once

You can also paste a file with Ctrl/Cmd+V, or from the clipboard menu.

Document languages
Pick up to 4. Fewer languages read faster and more accurately.

A scanned PDF has no text to extract, so the page is rendered as a picture, its words are recognized with Tesseract in the languages you pick, and rows are rebuilt from where each word sits. Words that share a baseline form a row; a wide gap starts a new cell. The grid then shows what was read so you can correct it before exporting.

Results depend on the scan. In our tests on clean, straight, synthetic 200 dpi pages, 97 to 98 percent of table cells came out right, but real documents with skew, shadows, stamps or small print do worse, and a single wrong digit in an amount is possible. Treat the output as a draft: check totals, dates and signs, and rescan at 300 dpi if the first read is poor.

Choose the language of the text, up to four, since recognition is better when it matches the page. Up to 50 pages that need recognition are handled per request on this site, and files can be up to 500 MB, so split a longer scan or enter a page range. Pages that already contain a text layer are read directly without OCR.

Features

  • OCR with Tesseract in up to four chosen languages
  • Rows rebuilt from word positions, with gaps marking cells
  • Editable grid before the Excel download
  • Plain numbers become numeric cells; codes stay text
  • Page range and a clear 50-page OCR limit per request

When to use this tool

  • Getting figures out of a scanned report or printed statement
  • Digitizing an old table that exists only as a scan
  • Drafting a spreadsheet from a photographed price list saved as PDF

How do I convert a scanned PDF to Excel?

  1. Upload the scanned PDF and enter the page range of the tables
  2. Choose the language or languages on the page
  3. Click Process and compare the grid with the scan
  4. Correct mistakes, then download the Excel workbook

Limits and edge cases

  • Up to 50 pages that need OCR per request on this site
  • Skewed, blurred or low-resolution scans reduce accuracy; verify every figure
  • Handwriting, merged cells and ruled borders are not interpreted

Test results and conditions

  • Test files: synthetic one-page PDFs generated for this project's test suite, not real customer documents. A scanned regional sales table (200 dpi, light blur and noise, 0.4 degree skew): 97.1 percent of cells exactly right. A scanned statement table: 98.5 percent.
  • The errors were single characters, for example Q1 read as Qi. Skewed, shadowed or small-print scans will do worse.

Privacy for this tool

Files are uploaded over HTTPS, processed in memory or a short-lived temporary directory on our servers, and deleted when your result is ready. We do not keep copies for later browsing, training, or advertising profiles. See the Privacy Policy for retention details and AdSense cookie disclosures.

Frequently asked questions

How accurate is OCR on tables?
In our tests on clean, straight, synthetic 200 dpi scans 97 to 98 percent of cells were right, but real scans vary a lot in quality. Always check amounts, dates and signs in the grid.
What resolution should I scan at?
300 dpi, black and white or grayscale, with the page straight. Higher resolution helps small print; skew and shadows hurt more than anything.
Can it read handwriting?
No. Recognition is built for printed text; handwritten tables will come out wrong or empty.
Why did it return fewer rows than I expected?
Rows are rebuilt from word positions, so faint text, very small print or low confidence words can be dropped. Rescan at a higher quality or add the missing rows in the grid.

Last updated:

Call this from code

Every tool on this site is a plain REST endpoint - no account or API key needed for anonymous use. Built for AI agents and developers as much as for browsers.

curl -X POST "https://pdf123.xyz/api/v1/convert/pdf/tables" \
  -F "[email protected]" \
  -F "pageNumbers=all" \
  -F "ocr=auto" \
  -F "languages=eng" \
  -o output.pdf

Also available as an MCP tool for agent clients that speak Model Context Protocol (JSON-RPC 2.0 over POST /mcp). Full API reference

Use it from an AI agent

Skill

Claude Code, Codex, Cursor and other AI agents can run Extract Tables from PDF for you with this skill: /skills/pdf123-pdf-to-tables.md

All agent skills

Files are used only for this processing job and deleted automatically afterward.