Invoice PDF → spreadsheet row
Drop an invoice PDF (or a photo, or paste the text). You get one clean row — vendor, invoice number, date, currency, tax, total — you fix anything that looks wrong, then download CSV that Excel and Google Sheets open natively.
Upload the PDF, take a photo, or paste the text. Text-layer PDFs are read automatically.
Every field is editable. Nothing is committed until you export, so a wrong tax line costs one click.
Keep adding invoices to the same table, then export the whole thing as one CSV.
Default columns (generic CSV):
vendor,invoice_number,date,currency,subtotal,tax,total,notes Corner Cafe,INV-2041,2026-03-14,USD,45.00,3.20,48.20,
There are also two accounting-shaped exports built from the same row: Xero bills and QuickBooks Online bills column sets. Due date, account code, and tax code are left blank on purpose — the tool will not guess your chart of accounts.
| Job | ReceiptSheet |
|---|---|
| One invoice → one summary row (vendor / number / date / tax / total) | Yes |
| Photos and screenshots of invoices | Yes |
| CSV export for Excel, Numbers, or Google Sheets | Yes |
| Line items, each product row from one invoice | No — one row per document |
| Batch PDF → single .xlsx workbook converter | No — you build the table, then export once |
| Ledger, reconciliation, VAT filing, approval flows | No — this is not accounting software |
| Reading page 2+ of a multi-page scanned PDF | No — the vision path reads page 1; paste the text for the rest |
Free: 3 documents/day without an account, 5/day signed in. Paid: $9/mo for 300 documents, $19/mo unlimited. No separate per-seat maths, no annual lock-in.
You process one document per extraction, but you can append each result to the same table and export the whole set as a single CSV. What it will not do is split one PDF containing many invoices into many rows — that needs line-item detection, which is out of scope.
No. It exports CSV. Excel and Google Sheets open CSV directly, and you can save as .xlsx from there. CSV also means the file works in Numbers, LibreOffice, and any import step in Xero or QuickBooks.
Text-layer PDFs are read automatically. For scanned, image-only PDFs you have two options: upload a photo of the page, or paste the invoice text. The vision path reads the first page only, so for a long scan paste the text instead.
No. Everything runs in the browser. An account is optional — it raises the free daily limit from 3 to 5 documents and keeps your plan across devices.
If AI extraction is enabled, the document text or image is sent to a model provider solely to propose field values. Results are not published anywhere and there is no public gallery. Server logs are kept for a limited period for security and support. Details are in the privacy policy.
Same job, different container: receipts to Excel follows the same one-document-one-row flow. Both are web-based, so there is nothing to install on a phone.