Sheet from PDF
No upload

PDF tables to Excel or CSV, without the upload

Drop in a PDF and get clean, editable rows out — bank statements, invoices, receipts, reports. Your browser does the reading, so nothing you convert touches a server, and there is no file size limit.

Drop a PDF with a table in it hereor click to choose a file — it stays on this device

How it works

Four steps, about ten seconds, nothing sent anywhere.

  1. Choose your PDFDropped straight into the page. It is opened locally and never sent anywhere.
  2. The layout is rebuiltThe text inside the PDF is grouped back into the rows and columns it came from.
  3. The columns are matchedHeaders are recognised as date, description, debit, credit and balance.
  4. Check, fix, exportCorrect any cell, then download CSV or paste straight into Excel.

Why the balance check matters

Most converters hand you whatever the PDF said and let you find the mistakes. This one adds the transactions up and compares the total against the balance column on every row. If a line does not reconcile, you are told which line — before it becomes a reconciliation problem three months later. It is a deliberately unglamorous feature, and it is the reason a bookkeeper can trust the output.

Nothing is uploaded

There is no upload step and no server-side processing. Turn off your internet after the page loads and the table is still read and the rows still built — that is how you can check for yourself. Only the paid export step needs a connection.

It reads the page layout, not just the text

Columns are found from where the text actually sits on the page, then matched to date, description, amount, debit, credit and balance — including tables that wrap a description over two lines. Bank statements get a dedicated layout reader on top of that.

Every row is checked

The transactions are added up and compared against the running balance, so a missing or misread line is caught before it reaches your accounts.

Questions

What people ask before trusting a tool with a bank statement.

Is my PDF uploaded to a server?
No. The PDF is opened and read by your own browser using pdf.js. There is no upload step and no copy of your file anywhere but your device. You can check that yourself: load the page, turn off your internet connection, and the table is still read and the rows still built. Only the paid export step needs a connection, because that is the part that checks the account.
Does it work with scanned PDFs?
Not yet. This tool reads the text layer inside a PDF. If your file is a scan or a photograph, there is no text layer to read and the extraction will come back empty. Scanned support needs optical character recognition, which is on the roadmap — the tool tells you when it thinks it is looking at a scan rather than guessing.
What if the columns are wrong?
You can remap every column yourself, and edit any cell in the grid before exporting. Extraction from real PDFs is imperfect, so the grid is editable by design rather than pretending the first attempt is always right.
Does it handle DD/MM/YYYY and MM/DD/YYYY?
It looks at every date in the file to work out which way round your document writes dates, and it tells you when the file is genuinely ambiguous. When it cannot tell, it asks rather than silently guessing — because a wrong date silently corrupts a set of accounts.