Free PDF table tool

Extract PDF tables without manual copying

Upload a report, invoice, statement, or research PDF. Edit the detected rows in your browser, then download a clean table for Excel, Sheets, or your next data task.

Open the free extractor
Selectable PDF text
Editable table preview
XLSX, CSV, TSV, JSON, and HTML export

A simple PDF → spreadsheet path

  1. 1. Select pages. Extract the whole file or a range such as 1-3,5.
  2. 2. Check the preview. Fix cells, remove repeated headers, and delete noisy rows.
  3. 3. Export. Download an XLSX workbook, CSV, TSV, JSON, or a standalone HTML table.

Try a real input

See the PDF and the extracted rows before you start

The sample is a one-page selectable-text invoice created for this extractor. Load it in one click or download it, then check the result against this preview.

Sample invoice table

One page, four columns, three line items, and selectable text. No account or email required.

Download sample PDF

Then drop the file into the extractor below and choose XLSX for Excel or CSV for Google Sheets.

Expected editable output

Item, quantity, unit price, and total stay in separate columns.

ItemQtyUnit priceTotal
Design review2$120.00$240.00
Data cleanup3$85.00$255.00
Final export1$60.00$60.00
PDF to ExcelPDF to CSVExtraction guide

Test missing cells and wrapped descriptions

These files are synthetic, not customer documents. They exercise repeated headers, blank cells, and continuation text. Single-cell lines now stay visible with a Review label, including document titles and notes. Remove unrelated notes only after comparing with the source. These fixtures are not an overall accuracy benchmark.

Two pages, repeated header, missing quantity

After reviewing and removing the fixture titles and notes, the expected table is one header plus Alpha, Beta and Gamma. Beta has a blank Qty: it must remain blank, not become zero. Gamma retains source page 2 in JSON.

Download repeated-header PDF
Item,Qty,Amount
Alpha,2,10.00
Beta,,20.00
Gamma,1,30.00

Wrapped description: retained for review

The continuation “includes two follow-ups” is kept as a separate Review row. Check the source, then use Merge into previous row to join it to Support package. The extractor does not guess whether a single-cell line is a continuation or a separate record.

Download wrapped-description PDF

Expected: header, Support package, continuation, and Export, plus reviewable fixture notes. Restore extracted rows can undo your edits.

Algorithm update: 2026-09-09. Multiple tables on one page, merged cells, scans and complex layouts still require manual review or another workflow. These two files are not a representative benchmark of real-world PDFs.

No signup required

PDF Table Extractor

Works best with selectable-text PDFs. Extraction and editing happen in your browser.

Free browser tool · no signup · no upload

Frequently Asked Questions

Can this extract tables from any PDF?
It works best with selectable-text PDFs where table rows have visible spacing. Scanned PDFs, photos, and heavily merged tables may need OCR or manual cleanup.
Can I try it before using my own file?
Yes. Click the sample invoice button in the extractor to load a one-page PDF locally, or download it and upload it yourself. Compare the editable preview with the expected rows shown here.
Can I choose only some pages?
Yes. Leave the page field blank for the whole PDF, or enter ranges like 1-3,5 to extract selected pages.
Can I fix extraction mistakes before downloading?
Yes. The preview is editable. Correct cells, remove noisy rows, and export the edited table.
Which formats can I download?
XLSX creates a real Excel workbook, and CSV or TSV open in Excel, Google Sheets, and Numbers. JSON is useful for code, and HTML creates a portable table report. Every cell is written as text, so IDs and leading zeros such as 0012 are preserved.
Is the extraction private?
Yes. The current extractor uses pdf.js in your browser. The PDF is not uploaded for this workflow.