fastdocai
Sign in

Scanned PDFs and photosreal tables

A scan has no text layer — there is nothing to copy, and selecting the table gives you nothing. This reads the page the way a person does, so a skewed scan or a phone photo still comes back as a spreadsheet you can sum.

  • An Excel file where the amounts are numbers you can sum — not text that looks like numbers.
  • A category on every row, and foreign spend converted at the rate on the day you spent it.
  • Or the file your software wants: OFX, a Xero or QuickBooks CSV, UBL XML for an ERP.

The PDF your bank sends

NORTHBANK PLC

Statement · 01–31 March 2026 · GBP

03/03/26SALARY - ACME LTD4,250.00

04/03/26RENT - MARCH1,200.00

05/03/26TESCO STORES 447187.35

07/03/26SHELL FUEL LONDON62.10

11/03/26AWS EMEA214.88

22/03/26VODAFONE LTD42.00

The Excel you get back

DateDescriptionCategoryGBPEUR

Category filled in for you. EUR converted at the rate on each transaction's own date — not today's.

Download asExcelCSVXMLOFXXero CSVQuickBooks CSVUBL XML

Drop a scanned PDF or a photo of a table

PDF · photo · iPhone HEIC · scan · zip · Word — mix them freely, no sign-up to try

Free to try, no sign-up. Pay $9 for a batch, or sign in for 10 free pages.

How it works

  1. 1

    Drop the scan

    A flatbed scan, a phone photo, a fax that became a PDF in 2009. All fine.

  2. 2

    It reads the layout

    Columns stay columns. A number in the debit column does not wander into credit.

  3. 3

    Take the table

    Excel, CSV or XML — with the amounts as numbers, not as text that looks like numbers.

No text layer needed

Traditional extractors look for text hiding under the image. When there is none — a scan, a photo — they return an empty table. This one looks at the page.

Crooked, creased, badly lit

A receipt photographed at a till is none of the things a parser wants. It is read anyway, because that is the form these documents actually arrive in.

Columns survive

OCR-then-parse flattens a page to a line of words and loses which column a number was in. The layout is read as a layout, so debit stays debit.

Questions people ask

Can it read a scanned or photographed document?
Yes. Scans and phone photos are read the way a person reads them — a crumpled receipt or a skewed scan still works, and there is no template to configure for each bank.
What if one PDF holds many receipts?
It splits them apart. Fifty receipts scanned into a single PDF become fifty rows; a twelve-page statement stays one document.
Are the amounts real numbers in Excel?
Yes. Amounts and balances arrive as numeric cells, so a column sums the moment you open the file.
How are foreign currencies handled?
Each amount is converted at the published rate for the date of that transaction — not today's rate. The rate and its date sit beside the figure so you can check it.
Do I need an account?
No. Convert a batch, pay once, download it. An account is only for a monthly page allowance.
Does it work with QuickBooks and Xero?
Yes. Alongside Excel, you can download a statement as OFX — the standard bank-data format both read — or as a ready-made CSV laid out the way each one expects: Date, Amount, Payee, Description, Reference for Xero, and the three-column Date, Description, Amount for QuickBooks, with money out already signed negative. Invoices and receipts can also come out as UBL XML for an ERP or an e-invoicing system.