Tiny Online Tools logoTiny Online ToolssearchSearch tools…grid_viewAll Tools
Homechevron_rightPDF Toolschevron_rightPDF Table ExtractorPDF Table Extractor

PDF Table Extractor

Find the tables inside a PDF and export them to CSV or XLSX, with the detection method and confidence shown for every table.

upload_file

Click to browse or drag & drop files here

Select a PDF to scan for tables

Accepted: .pdf,application/pdf

Similar Tools

PDF Embedded File Extractor

PDF Embedded File Extractor

Find and download every file attached inside a PDF, including attachments hidden in annotations and PDF 2.0 associated files.

PDF Link Extractor

PDF Link Extractor

List every link in a PDF — external, internal, mailto, and the bare URLs in the text that carry no annotation — with pages, anchor text and risk flags.

PDF Structured Data Extractor

PDF Structured Data Extractor

Turn a PDF into structured JSON — headings, paragraphs, lists, tables and images — from its tagged structure tree or from layout inference.

PDF Image Extractor

PDF Image Extractor

Pull the original embedded images out of a PDF — JPEGs byte for byte, everything else as PNG — with intrinsic size, drawn size and effective DPI.

PDF Font Inspector

PDF Font Inspector

List every font in a PDF with its subset tag, subtype, embedding status and the pages that actually use it.

CSS Minifier

CSS Minifier

Minify CSS by removing whitespace, comments, and redundant rules.

Advanced Dice Roller

Advanced Dice Roller

Roll dice with full notation support: XdY+Z, keep highest/lowest, exploding dice.

apps

More Tools

Browse our full collection of free online tools.

Getting a table out of a PDF

A PDF has no tables in it. What it has is text placed at coordinates, and sometimes lines drawn around that text. A table is something you see when you look at the page. Any tool that claims to "read the tables" is guessing — the only honest question is how good the guess is, and this tool tells you.

Two ways to guess

Ruled detection uses the borders the producer actually drew. The horizontal and vertical rules are clustered into a grid and every word is dropped into the cell it falls in. When a document has borders this is close to exact, which is why these results are reported at high confidence.

Whitespace detection is for the majority of real documents, which have no borders at all. The page is swept for vertical gaps that stay blank across many consecutive lines; those gutters become the column boundaries. It works well on financial statements, price lists and reports, and it can be fooled by a two-column paragraph.

Leave the method on Auto and both passes run; the stronger, non-overlapping results are kept.

Tuning the result

  • Minimum rows and columns filter out noise. A two-row, two-column "table" is very often a heading and a date.
  • Pages accepts all, 1-3, or 2,5,9. Scanning one page of a 400-page report is much faster than scanning all of it.
  • Every detected table can be deleted before export. Detection is a heuristic; if it found a paragraph, remove it rather than cleaning up the spreadsheet afterwards.

Export

  • One CSV per table, or all of them zipped together.
  • One combined CSV, with a marker row between tables so the boundaries survive.
  • One XLSX with one worksheet per table, ready to open in Excel, LibreOffice, Numbers or Google Sheets. Cells that are plain numbers are written as numbers so they sum correctly; anything that merely looks numeric, such as an invoice code with a leading zero, stays text.

Every CSV carries a UTF-8 byte-order mark, so accented characters survive the trip into Excel on Windows.

Privacy

The PDF is opened, parsed and exported entirely inside your browser. Nothing is uploaded, and the tool makes no network requests while it works.