Upload a Chinese VAT invoice (增值税发票 / fapiao), train ticket, flight itinerary or taxi receipt, and get structured fields back as data — invoice numbers, tax IDs, line items and totals — not lines of raw text. Built for developers integrating China document recognition. No sign-up.
Privacy: your file is processed in memory only — never written to disk, never stored, never kept after the result is returned.
Four Chinese document types, classified automatically: VAT invoices （增值税发票， including electronic fapiao), train tickets, flight itineraries （行程单） and taxi receipts. Each type returns its own typed fields — invoice code and number, buyer and seller names and tax IDs, amounts and totals for invoices; train number, stations, date, seat and passenger for tickets.
Generic OCR returns lines of raw text and leaves the parsing to you. This service classifies the document first, then extracts named, typed fields as structured data (JSON) — totals come back as numbers, dates as dates, ready to store or validate.
The free demo recognizes one document at a time: up to 10 documents per day per visitor, files up to 2 MB, images or PDF. A typical document takes about 1–2 seconds. The API removes the daily cap.
Yes — through the API. Its batch endpoint accepts up to 10 documents in a single call and returns a result per document. The demo here is intentionally single-document so you can judge recognition quality first.
No. Files are processed in memory and discarded as soon as the result comes back — never written to disk, never retained, never used for training or anything else. We only keep a daily counter keyed by a hashed IP to enforce the free demo limit.
Free demo · 10 documents per day · Limits reset daily