Invoice extraction

Invoice PDF to Excel: one workbook per supplier, source embedded.

Flattening invoices into a CSV loses the things that matter — the line items, the VAT split, the link back to the document. Pileform keeps all three: lines extracted with their rates, grouped one workbook per supplier, original page embedded beside the rows.

The invoice problem

A quarter of supplier invoices means hours of copy-paste: invoice number, date, net, VAT, gross — times every line, times every page. Generic converters flatten the document into a table that loses line-level VAT and keeps no connection to the source. Then someone asks 'where did this figure come from?' and the answer is a folder dig.

Pileform extracts every line with its VAT, groups invoices by supplier, and embeds the original page next to its rows — the answer to 'where did this come from?' is one click.

How it works

Many invoices, one PDF, one drop.

Combine the period's invoices into one PDF — order, orientation, and page boundaries do not need to be tidy. Extraction finds each invoice, reads its lines, and files it under its supplier.

Line items, kept.

Each invoice line comes out as its own row — description, quantity, net, VAT rate, VAT amount, gross. Totals reconcile to the invoice's printed totals, with rounding gaps in explicit adjustment cells.

Hosting — March · €40.00 · 19% · €7.60 and Domain renewal · €12.00 · 19% · €2.28 as separate rows, reconciling to the invoice total of €61.88.

VAT per line, never per document.

Mixed-rate invoices keep each line at its printed rate — including reverse-charge candidates, which are flagged and routed through reverse-charge handling instead of defaulting to ordinary purchases.

One workbook per supplier.

Forty invoices from one supplier consolidate into a single tab with a running total and the VAT control figures. The period stops being a pile of PDFs and becomes a set of supplier accounts.

The source page, embedded.

Every row links to the original invoice image, embedded inside the workbook itself. Audit queries, client questions, and inspector requests are answered from the file — years later, without a document archive dig.

Post-ready for Xero and QuickBooks.

Confirmed entries post straight through — supplier invoices as Bills in Xero, journal entries in QuickBooks — with per-line VAT carried over. The Excel and CSV exports are always yours.

See an invoice become rows.

The sample workbook on the features page shows the per-supplier tabs, the line-level VAT, and the embedded source pages.

See the sample workbook
Honest answers

Invoice extraction questions, answered.

Line items. Each invoice line becomes its own row with description, quantity, net, VAT rate, VAT amount, and gross — and the rows reconcile back to the invoice's printed totals. Flattening to one total per invoice loses the VAT split, which is exactly what the return needs.

Yes — that is the intended workflow. Combine the period's invoices into one PDF in any order; extraction detects where each invoice starts and ends, reads it, and files it under its supplier. A typical 200-page batch processes in about 12 minutes.

The rate printed against each line is read and applied — 55 jurisdictions' rates are in the engine, and a printed rate is never overridden. Missing rates are inferred and yellow-flagged for review, and reverse-charge candidates (foreign suppliers, construction services) are flagged and routed through reverse-charge handling.

The gap is surfaced, not hidden. Cash rounding shows in an explicit adjustment cell; anything beyond rounding flags the document for review with the source page right there. Nothing is silently fudged to force the total to tie up.

Yes. Confirmed entries post as Bills in Xero and journal entries in QuickBooks, carrying the per-line VAT. Everything also exports to Excel and CSV regardless, so the records stay yours whatever ledger you use.

Stop copy-pasting invoices.

Sign up free, no card. Drop a real period's invoices and see the lines, the VAT, and the embedded sources before you commit to anything.