본문으로 건너뛰기
PPDF123

Extract Tables From a PDF

Extract Tables finds aligned table rows in a PDF, shows each table in an editable preview, and exports your corrected rows as Excel or CSV. Pages without text can be read with OCR, so check numbers read from scans before relying on them.

여기에 파일을 끌어다 놓거나, 클릭해 고릅니다

PDF 지원 · 파일당 최대 500.0 MB · 한 번에 최대 20개 파일

Ctrl/Cmd+V 또는 클립보드 메뉴로 파일을 붙여넣을 수도 있습니다.

문서 언어
최대 4개까지 선택하세요. 언어가 적을수록 더 빠르고 정확하게 읽습니다.

Extract Tables finds table-like blocks in a PDF and shows them in a preview you can correct before anything is downloaded. Rows are detected from aligned columns in the PDF's text layer, and columns follow the blank channels the rows share, so a cell left empty keeps its place instead of shifting the rest of the row.

In the preview, click a cell to fix it, delete a stray row or column, merge a wrapped line into the row above, or untick a table you do not want. Excel and CSV downloads use exactly what the preview shows. In Excel, plain numbers such as 1,234.50 become numeric cells, while values with leading zeros, symbols or units stay text.

Scanned pages are optional. With OCR on, pages that have no text layer are rendered, recognized with Tesseract in the languages you choose, and rebuilt into rows from the word positions. That works best on clean, straight scans; check amounts and dates read from a scan before you rely on them. Up to 50 pages that need OCR are handled per request on this site.

기능

  • Editable preview: fix cells, delete rows or columns, merge wrapped lines
  • Excel and CSV exports built from the corrected preview
  • Empty cells keep their column; plain numbers become numeric cells in Excel
  • Optional OCR for scanned pages, with a language choice
  • Optional page range

이런 경우에

  • Moving a bank or card statement into a spreadsheet
  • Extracting an invoice's line items or a report's tables
  • Checking what a PDF's tables look like before choosing an export

How do I extract tables from a PDF?

  1. Upload the PDF and, if needed, enter a page range
  2. Leave OCR on for scanned pages and pick their languages, or turn it off to read only the text layer
  3. Click Process and review each table in the preview
  4. Correct anything that is off, then download Excel or CSV

제한과 경계

  • Detection needs aligned columns; tables typeset with single spaces between columns may not register
  • Merged cells, colored headers and borders are not reconstructed
  • OCR on scans is less exact than reading a text layer: verify figures, dates and signs
  • At most 50 pages that need OCR per request on this site
  • Upload limit on this website: 500 MB per file, sent in chunks above about 95 MB. A single direct API request body is capped at 100 MB.

자주 묻는 질문

Why were no tables found?
A table needs at least two rows with aligned columns. Prose, a single row and tables with single-space gaps do not qualify. For image-only pages, keep OCR on.
How accurate is OCR on a scanned table?
It depends on the scan. Clean, straight, high-resolution pages usually come out well, but a wrong digit or a shifted column is possible, so check the preview before exporting.
Why is a number stored as text in Excel?
Only plain numbers become numeric cells. Values with a leading zero, a currency symbol, a percent sign or a unit stay text so account numbers and codes are not altered.
Does the preview change my PDF?
No. Edits only affect the Excel or CSV you download; the PDF is left alone.

최종 업데이트:

코드에서 호출하기

이 사이트의 도구는 모두 일반 REST 엔드포인트입니다. 익명 사용에는 계정이나 API 키가 필요 없습니다. 브라우저뿐 아니라 개발자와 AI 에이전트도 그대로 쓸 수 있습니다.

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

Model Context Protocol을 쓰는 에이전트 클라이언트용 MCP 도구로도 제공합니다(JSON-RPC 2.0, POST /mcp). 전체 API 문서

AI 에이전트에서 사용하기

스킬

이 스킬을 설치하면 Claude Code, Codex, Cursor 등 AI 에이전트가 'PDF에서 표 추출' 작업을 대신 실행합니다: /skills/pdf123-pdf-to-tables.md

전체 에이전트 스킬

파일은 이번 처리에만 쓰이며, 끝나면 자동으로 삭제됩니다.