Technical explainer

How receipt data extraction works

Extraction is not one step. The system must turn a file into readable text, identify candidate fields, reject ambiguous values, let a person correct the result, and format the reviewed data.

The five-stage pipeline

  1. 1

    Input

    Photo or PDF

  2. 2

    OCR

    Pixels to text

  3. 3

    Parsing

    Text to fields

  4. 4

    Review

    Human correction

  5. 5

    Export

    Fields to CSV

FieldStrong evidenceCommon ambiguity
VendorDistinct merchant headerGeneric headings, email wrappers
DateExplicit printed purchase dateOrder, ship, and print dates together
TotalAmount tied to Total/Amount DueSubtotal, tip, balance, or card amount
TaxTax/GST/VAT label and amountIncluded tax or no printed tax line

Conservative extraction

A blank field is visible and correctable. A confident-looking wrong value can silently contaminate totals. ReceiptJar therefore leaves uncertain values blank rather than always filling every column.

Confidence is not correctness

Confidence reflects available evidence, not a guarantee. Clear receipts can still contain unusual layouts, and poor images can occasionally produce plausible but wrong text.

How to evaluate a receipt extractor

See the public regression setExplore the review flowReceipt-to-CSV workflow