InvoiceXL / Practical guide

Invoice PDF to Excel with an existing spreadsheet or Tabula

Try a local text-table workflow, then repair invoice page boundaries and validate the worksheet before relying on it.

Try a small local extraction

Before moving an invoice PDF into Excel, check whether its text is selectable and whether the supplier can send the original spreadsheet. If your Excel installation includes From PDF, open a copy through Data, Get Data, From File, From PDF. Inspect candidate tables in Navigator and transform the relevant item table before loading. If this connector is unavailable, Tabula can extract a selected text-PDF region to CSV for free. Import that CSV into your spreadsheet with explicit data types. These methods may avoid a separate conversion purchase for a simple document.

Capture headers separately from the item table

The invoice number, supplier, date and currency may appear outside the extracted table. Enter these in a document register after checking the labels in the PDF. Do not assume that the first date is the invoice date or that a number near the top is the invoice number. Keep delivery notes, purchase-order references and payment instructions separate. If multiple document dates matter, give them distinct column names. A copied item table is only part of the invoice, so an extraction that looks complete can still lack essential header evidence.

Repair page breaks before calculating

A multi-page invoice may repeat item headings and carry a subtotal forward. Compare the final item on each page with the first on the next. A description can wrap without creating another billable line. Remove repeated headers and page carry-forwards only after checking their meaning. Keep a source-page field on the working table so edits remain traceable. If separate tables have different layouts, align them by meaning before appending. A quantity and a unit price shifting one column can still produce visually plausible rows while changing the interpretation of the document.

Assign data types deliberately

Mark invoice numbers, product codes and purchase-order references as Text before automatic conversion removes leading zeros. Select a number locale that matches the printed punctuation. A date such as 04/05 requires source context; retaining the printed value avoids losing evidence during cleanup. Check parentheses and minus signs on discounts or credits. Keep blanks distinct from zero. If a value cannot be read from the PDF, mark it unresolved instead of calculating a replacement that merely makes the total match. Preserve the raw extraction sheet before applying these transformations.

Verify the finished local workbook

Count items on every page, compare a short and long description, and inspect each quantity, price and line amount. Check subtotal, tax, shipping and discount at the document level. Do not add tax twice when printed item amounts already include it. Keep document totals once per invoice, outside the item sum. If the PDF is a scan, a text-table extractor alone is insufficient; obtain a clearer digital source, use a suitable OCR workflow or transcribe carefully. A polished worksheet is useful only after these source checks, regardless of which extraction method produced it.

Create a reproducible local extraction sequence

Keep the source PDF and name the raw extraction sheet before making changes. Select only the item region in Tabula, or inspect the candidate table in Excel's Navigator where From PDF is available. Export or load a copy and record which source pages it covers. Inspect the first and final item on each page before appending tables. If the next page changes column order, rename and align fields before combining them. Keep a small transformation note describing removed headings, joined descriptions and converted numeric columns so a reviewer can follow the cleanup without guessing from the final worksheet.

Check a quantity-price shift using one row

For a printed item showing quantity 4, unit price 12.50 and amount 50.00, verify that the extracted cells keep those three roles. A shifted column can place 12.50 under Quantity while still looking like a valid number. Compare the source labels and recompute the row only after that alignment check. If the invoice uses a discount, unit conversion or inclusive tax, do not force a simple quantity-times-price formula to agree by changing the extracted values. Mark the row for interpretation and retain the printed amount. Arithmetic is a useful diagnostic after the table structure has been verified.

Where InvoiceXL fits

InvoiceXL is another extraction route when a local table workflow is insufficient. It supports one invoice per PDF, up to 20 pages, using local English OCR for scans and retaining source filename/page references on line items. It does not guarantee recognition of handwriting, rotated or faint scans, inclusive tax, repeated tax summaries or complex discounts; these may fail or require checking. One invoice is free with an account and no card. Paid batches take up to ten PDFs, 15 MiB combined and 3,000 line items. A successful PDF, including usable flagged output, costs one credit. Files expire within 24 hours.

Try a real free conversion See one-time packs

Written by Zijian Xiao. Examples are synthetic. Review all extracted results against your original source file.