Skip to content
PDFBetter
Convert from PDFBeta

Convert invoices to data, with every total checked

Drop in up to ten invoices, fix anything the math check flags, and download the data as Excel, CSV or a QuickBooks or Xero bill import.

Drop your files here

or paste with Ctrl/⌘+V

Uses AI (OpenAI, not used for training). 5 free AI credits without an account, 20 a month on the free plan.

PDF · Word · Excel · PowerPoint · JPG · PNG · up to 10 MB

Try it 5× without an account. A free account includes 20 AI credits a month. See plans

What it does

Invoice converter, explained

To convert an invoice into data, drop the PDF (or a scan or photo of it) on this page and choose Read. PDFBetter's AI reads the supplier, bill-to name, invoice number, invoice and due dates, PO number, currency, every line item (description, quantity, unit price, amount), subtotal, discount, shipping, tax and total. Each value keeps the page and the exact words it was read from, and values whose quote can't be found on the page are marked “check”. It costs from 1 AI credit per page read.

Read more

Then the numbers are checked with plain arithmetic, not AI: each line's quantity × unit price must equal its amount, the line amounts must add up to the subtotal, and subtotal − discount + shipping + tax must equal the total, to half a cent (or half a yen for currencies without cents). If one line was misread, that line is highlighted with the amount it's off by. You can correct any field or line in the browser and the check reruns as you type.

When it all adds up, export an Excel workbook (Invoices, Line items and Checks sheets), a CSV with one row per invoice or per line, or a bill import file for QuickBooks Online or Xero. Exports are built in your browser from what's on screen, including your edits, and don't use credits.

An invoice parser that only extracts fields gives you a spreadsheet that looks right even when a 7 was read as a 1. Here every invoice carries a visible check, so the ones that need a second look stand out before they reach your books. For your own invoice formats or other documents, PDF to JSON lets you define any fields you like.

How to do it

Convert invoices to data, with every total checked in 4 steps

  1. 1Your PDF

  2. 2Convert

  3. 3The new format

  1. 1

    Add invoices

    Up to 10 at once: digital PDFs, scanned PDFs or photos (JPG, PNG, HEIC). Up to 10 MB each on the Free plan, 50 MB on Starter.

  2. 2

    Read them

    A free account is needed. Each file uses from 1 AI credit per page read; photos and scans get a text layer first.

  3. 3

    Check the math

    Each invoice shows whether its lines, subtotal and total agree. A red line tells you where and by how much.

  4. 4

    Fix and export

    Edit any value, add or delete lines, then download Excel, CSV, or a QuickBooks or Xero bill file.

Example

Before and after

Before

A one-page supplier invoice with two line items (2 × 50.00 paper, 1 × 50.00 toner), Subtotal 150.00, Shipping 10.00, Tax 16.00 and Total 176.00.

After

One invoice row with the supplier, number, dates and 176.00 total; two line rows; and a Checks sheet showing each line, lines = subtotal, and 150.00 + 10.00 + 16.00 = 176.00, all OK.

Settings & formats

What you can control

Fields read
Supplier, bill to, invoice number, invoice date, due date, PO number, currency, subtotal, discount, shipping, tax, total and line items.
Math check
qty × unit price = amount per line; lines = subtotal; subtotal − discount + shipping + tax = total. Half a minor unit of tolerance.
Evidence
Every value keeps its page and quote; values whose quote doesn't match the page are flagged to check.
Exports
Excel (.xlsx), CSV (per invoice or per line), QuickBooks Online bills CSV and Xero bills CSV.

Input

PDF (digital or scanned)JPGPNGHEIC

Output

XLSXCSVQuickBooks bills CSVXero bills CSV

Good to know before you start

  • Up to 10 invoices per run, and each file is read as one invoice. A PDF holding several invoices should be split first.
  • On the Free plan the first 10 pages of each invoice are read (50 on Pro); pages beyond that are listed as not read.
  • The check proves the printed numbers agree with each other, not that the invoice is right: a wrong price that is carried through consistently still adds up. Tax rates and supplier tax numbers aren't validated.
  • When an invoice shows payments or deposits, the check uses the invoice total, not the balance due.
  • Handwritten amounts, stamps and very faint scans may not be read; type the missing values in.
If something goes wrong
An invoice doesn't add up
Use “show” next to the failing check: it scrolls to the line or total. Usually one digit was misread or a subtotal line was read as an item; fix or delete it and the check turns green.
“No text could be recognised”
The scan or photo is too faint, blurred or skewed for text recognition. Rescan at 300 dpi or take a flatter, brighter photo.
A line has no quantity or unit price
Many invoices print only the amount. The line is then counted in the subtotal check but has no line check of its own.

Invoice converter: questions people ask

What does the invoice parser extract?

The supplier, bill-to name, invoice number, invoice and due dates, PO number, currency, subtotal, discount, shipping, tax, total, and a table of line items with description, quantity, unit price and amount.

How is this different from PDF to JSON?

PDF to JSON lets you define any fields and returns them with evidence. This tool uses fixed invoice fields so it can add them up: it checks every line and total and exports in invoice and bill layouts.

Does it work with scanned or photographed invoices?

Yes. Photos become a one-page PDF in your browser, and pages without a text layer get one through text recognition before the AI reads them. That uses OCR pages as well as AI credits.

What if the discount is already taken off the subtotal?

The check tries the layouts invoices actually use: discount before or after the subtotal, shipping inside or outside it, and line prices with or without tax. It passes only on an exact match and says which reading matched.

What does it cost?

Reading uses from 1 AI credit per page read (the Free plan includes 20 AI credits a month, Pro 500); scans and photos also use OCR pages (20 a month on Free). Exports are free and can be repeated.

Where does my invoice go?

Page text is sent to our AI provider, OpenAI, under API terms that exclude it from model training. The files are saved to your PDFBetter documents, where you can delete them.

Which currencies work?

Any. The check uses the currency's minor unit (two decimals for most, none for yen, three for Kuwaiti dinar), and amounts are never converted between currencies.