Which invoices are supported?

Mainly Chinese e-invoice PDFs with a text layer: VAT electronic ordinary and special invoices, and the fully digital invoices introduced from 2023. OFD files need converting to PDF first. Scans of paper invoices can be read with the OCR option, though accuracy depends on scan quality. For English invoices only basic fields — invoice number, date and totals — are extracted.

What do the colours mean, and what if something is wrong?

Green means cross-checked (matches the QR code or passes the amount and check-digit tests), white means a valid format without cross-checking, yellow means please check (for example the QR code disagrees with the text), and red means not found. You can edit any cell; edited cells turn blue and the CSV export and renaming use your values.

What is the QR code used for?

The QR code in the top-left corner of a Chinese e-invoice holds the invoice type, code, number, amount, date and check code. The tool renders the page, decodes it and compares each item with the text it found: matches raise the confidence, mismatches are flagged for you to check, and missing values are filled from the QR code. Decoding happens locally; nothing is looked up online.

Are my invoices uploaded? Can it verify them?

Nothing is uploaded — all reading happens in your browser. The tool only extracts content and checks that amounts and formats are consistent; it can’t tell whether an invoice is genuine and isn’t tax or expense advice. Use the official State Taxation Administration verification service to check authenticity.

🧧

🧧Chinese E-Invoice Extractor - Batch PDF Invoice Data to CSV

Copying invoice numbers, amounts and tax IDs from a stack of Chinese e-invoice (fapiao) PDFs by hand is slow. Drop them all in and the tool reads each PDF’s text layer in your browser, using text positions to find the value next to each label — invoice code, invoice number, date, check code, buyer and seller names and tax IDs. It handles both the older VAT e-invoice layout (buyer above seller, labels and values drawn separately) and the newer fully digital invoice (buyer and seller side by side, 20-digit number). It also picks up the amount and tax from the total row, the total in figures and in Chinese words, remarks, and the line items (name, spec, unit, quantity, price, amount, rate and tax, joining names that wrap). With QR reading on, page 1 is rendered and the invoice QR code (01,type,code,number,amount,date,check code…) is decoded and compared field by field with the printed text; the tool also checks that amount + tax = total, that the amount in words matches the figures, and the check digit of 18-character tax IDs. Each field is coloured by confidence — cross-checked, valid format, please check, not found — and every cell can be edited. Totals exclude duplicates, which are flagged when the invoice number repeats. Export a UTF-8 CSV with BOM that opens cleanly in Excel (long numbers stay intact), a line-item CSV, or a ZIP of the PDFs renamed with a pattern such as {date}_{seller}_{total}. Scans can be read with local OCR, and a basic English invoice mode picks out number, date and total. Everything stays on your device; results are for organising only and are not tax advice.

Free · No sign-up · Runs in your browserLast updated: 2026-10-26

Loading tool...

How to use Chinese E-Invoice Extractor

1Select or drop several invoice PDFs; each one is read automatically
2Review the colour-coded fields, open the details for line items and checks, and fix anything that’s wrong
3Export the CSV summary or line items, or download the PDFs renamed with your pattern

FAQ

Which invoices are supported?: Mainly Chinese e-invoice PDFs with a text layer: VAT electronic ordinary and special invoices, and the fully digital invoices introduced from 2023. OFD files need converting to PDF first. Scans of paper invoices can be read with the OCR option, though accuracy depends on scan quality. For English invoices only basic fields — invoice number, date and totals — are extracted.
What do the colours mean, and what if something is wrong?: Green means cross-checked (matches the QR code or passes the amount and check-digit tests), white means a valid format without cross-checking, yellow means please check (for example the QR code disagrees with the text), and red means not found. You can edit any cell; edited cells turn blue and the CSV export and renaming use your values.
What is the QR code used for?: The QR code in the top-left corner of a Chinese e-invoice holds the invoice type, code, number, amount, date and check code. The tool renders the page, decodes it and compares each item with the text it found: matches raise the confidence, mismatches are flagged for you to check, and missing values are filled from the QR code. Decoding happens locally; nothing is looked up online.
Are my invoices uploaded? Can it verify them?: Nothing is uploaded — all reading happens in your browser. The tool only extracts content and checks that amounts and formats are consistent; it can’t tell whether an invoice is genuine and isn’t tax or expense advice. Use the official State Taxation Administration verification service to check authenticity.