PDF to Excel

Get the tables out. Check them first.

Drop a PDF and PDFMod finds its tables and draws each one over the page — columns and all — so you can see what it found before a single cell is written. $99 once and it is yours, and we never see your files.

You see and confirm every table — nothing is guessed into a spreadsheet behind your back

Never think about a PDF subscription again

Acrobat is about $155 a year and Smallpdf about $108, and both stop working the day you stop paying. PDFMod is $99, once, and the whole toolkit is yours — this, editing, redaction, signing, OCR, all of it. Pull a table out of a statement in four years’ time without renewing anything. If you only need it this week, it is $25 for the week.

And your bank statement, your invoice, your price list stay on your laptop. Not “encrypted in transit”, not “deleted after an hour”: we never see them, because nothing you open here is stored on a server. There is no account to make before you try it — drop a file in and watch it find your tables right now.

What you get

• Tables you can actually work with. A real .xlsx that opens in Excel, Numbers, LibreOffice and Google Sheets — sort it, filter it, sum a column. One sheet per table by default, each named for the page it came from, or one sheet per page if you’d rather.
• You see it before it is written. Every table PDFMod finds is outlined on the page with its column splits shown, and its cells appear in a preview. Drag a split to move it. Switch off anything that isn’t really a table. Then, and only then, save.
• It finds tables three ways.From the file’s own table tags, from ruled borders, and from columns lined up by spacing across several rows — so a bordered invoice and a borderless report both come out.
• Your original is untouched. PDFMod writes a new spreadsheet and hands it to you. The PDF you started from is exactly as it was.

What it does not pretend to do

Turning a page of glyphs back into a grid is genuinely hard, and the honest thing is to tell you where it is approximate rather than hand you a spreadsheet that looks finished and is subtly wrong.

• Rows are text lines. A cell whose text wraps onto two lines comes out as two rows. You can see that in the preview before you save, and split or rejoin in your spreadsheet after.
• Merged and spanning cells are flattened. A header that spans three columns lands in one of them, and a short row is padded on the right rather than guessed. The table says so, on the page, before you export it.
• It will not invent a table. If it cannot find a grid it is confident about, it says so and points you at PDF → Text — rather than hand you one word per row and call it a table.
• A scan has no text. A photographed page has no numbers in the file to extract. Run it through OCR first, on your machine, and then its table becomes real text.
• Sideways text is left alone. This reads upright, left-to-right tables. A table rotated on the page is skipped rather than scrambled.

If you would rather check than trust us

“We never see your files” is a claim, and you should not have to take it on faith. So it is enforced rather than promised, and you can watch it before you pay.

• Watch it yourself.Open your browser’s network tab, drop a statement in above, and pull its tables out. Nothing goes out. Finding the tables is free precisely so you can run a real job and see that before you decide.
• The browser enforces it, not our good intentions. This page is served with a Content-Security-Policy whoseconnect-src names only its own origin, so the browser refuses any request anywhere else — no analytics, no error reporting, no CDN. curl -I the page and read the header for yourself.
• Everything runs in the tab. Your file is read, its tables found, and the spreadsheet written by code already on your machine — which is also why there is a size ceiling, since a huge document runs the tab out of memory.
• Web browser only, this release. There is no Mac app yet. When one ships it will be signed, and the same policy will be written into it.

Questions people actually ask

Where does my document go?

Nowhere. It is opened, read and turned into a spreadsheet in this tab, on your machine. We never see it — there is no server here to send it to, and the page is served with a Content-Security-Policy that only permits connections back to its own origin, so the browser physically refuses to let it talk anywhere else. Open your network tab and pull a table out; you will watch nothing leave.

How do I know the table came out right?

Because you see it before it is written. Every table PDFMod finds is outlined on the page, with the column splits drawn in, and its actual cells are shown in a preview beside it. If a column split landed in the wrong place, drag it and the preview updates. If something isn't really a table, switch it off. Nothing becomes a spreadsheet until you have looked at it — a table that is quietly wrong is the one failure this tool is built to prevent.

What kinds of table can it find?

Three. A table the PDF tags as one — common in files from Word and Google Docs — comes across with its cells already grouped. A table drawn with visible borders is read from its rules. And a borderless table, whose columns are only lined up by spacing, is found when that spacing recurs across several rows. What it will not do is guess: a couple of coincidentally-aligned words is not a table, and it would rather find nothing than invent a grid.

Does each cell come out exactly?

The text does. The structure is approximate, and the page says so rather than pretending. Rows are text lines, so a cell whose text wraps onto two lines becomes two rows — you can see that in the preview. Merged cells and headers that span several columns are flattened, and a row with fewer cells than its neighbours is padded rather than guessed. This is honest table extraction, not a promise of perfect structure — that promise is the category's standard lie.

One sheet per table, or one big sheet?

One sheet per table by default, each named for the page and table it came from — so “the totals on page four” is a tab that says exactly that. There is a switch to put every table on a page onto one sheet instead, stacked with a blank row between them, if that is how you'd rather work with them.

My PDF is a scan. Can I get its table?

Not until it has text in it. A scan is a photograph of a page — there are no words or numbers in the file to pull into cells. Run it through OCR first, which is in this app and also runs on your machine, and the recognised text becomes real content you can extract.

It didn't find a table that I can see on the page.

Two usual reasons. The table may be set sideways, which this tool does not read — it works on upright, left-to-right text. Or the “table” may be a picture pasted into the PDF, in which case it is a scan of a table and wants OCR first. If it has text and simply isn't laid out as a grid, PDF → Text will get the words out plainly.

Need something else? All of PDFMod’s tools work the same way: on your machine, and we never see your files.