Client-side PDF table extraction
Pull every table out of a PDF. Watch it happen.
Quarry extracts tables from a PDF to CSV that Excel opens cleanly, free and entirely in your browser. Drop in a dense report and translucent boxes snap onto every table region it found, before you click anything. One click turns the box you actually wanted into clean, sortable CSV.
Nothing is uploaded. Your document never leaves the tab.
-
01
Drop it in
The page renders and detection runs immediately. No upload, no queue, no account. Ruling lines and whitespace both feed the same pass.
-
02
See what it found
Every candidate region is drawn on the page itself, dashed where the detector was unsure. You are never guessing which box became which table.
-
03
Fix it by hand
Drag an edge to recover a clipped column, discard a false positive, sort the preview, then export exactly what you see as CSV.
Getting a table out of a PDF without handing it to a stranger
Most tools that extract a table from a PDF to Excel for free want you to upload the document first. That is a bad trade when the file is an unreleased filing, a client's numbers, or anything covered by an NDA. Quarry does the whole job inside the browser tab: the PDF is read with pdf.js, analysed, and turned into CSV without a byte of it leaving your machine. Load the page once and it keeps working with the network off.
It reads the document's geometry, not a picture of it
As a pdf table to csv converter online, Quarry works from what the file already knows about itself. Every text run in a PDF carries an exact position on the page, and most report tables are drawn with real ruling lines. Quarry groups runs into rows by vertical overlap, finds column boundaries by projecting whitespace down the block, and snaps a region's edges to the ruling lines when the table has them. Tables drawn without lines fall back to their whitespace bounds. No model is guessing, so the same document gives the same answer every time, and every decision is one you can see on the page.
When it gets a boundary wrong, you can see it and fix it
Every extractor misjudges an edge sometimes. Most of them fail invisibly: you pull table data out of a PDF, open the CSV, and find a column missing with no clue where it went. Quarry draws each region it found onto the rendered page, dashed where it was unsure, so a bad boundary is something you catch before you export rather than after. Drag the edge to take back the column it clipped, discard a region that was never a table, sort the preview to check the numbers look sane, then export exactly what you see.
Questions people actually ask
- How do I extract a table from a PDF to Excel for free?
- Open the app, drop the PDF in, click the region you want, and press Export CSV. The file is written with a UTF-8 byte order mark, so Excel opens accented characters and currency symbols correctly instead of mangling them. No account, no upload, no paid tier.
- Is there a pdf table to csv converter online that does not upload my file?
- This one. Quarry is a static page with no server to send anything to. You can check that claim yourself: open your browser's network panel while you extract, or pull the network cable once the page has loaded and carry on working.
- How do I pull table data out of a PDF when the boundary is wrong?
- Drag the handles on the region box. The preview and the exported CSV both follow the box, so a column clipped off the right edge comes back the moment you widen it. Regions that were never tables can be discarded from the list.
- Does it work on scanned PDFs?
- No. Quarry reads the text and vector lines a PDF already contains, and a scan is an image of a page with no text layer underneath. The rule of thumb: if you can select the text in your PDF reader, Quarry can read it. If you cannot, run OCR first and then come back.
- Which browsers does it work in?
- It runs anywhere pdf.js and canvas do, which covers current Chrome, Edge, Firefox, and Safari on both desktop and phone. The automated browser suite runs against Chromium. No extension or plugin is needed.