Loading page
About this tool
Compare observable invoice fields and payment details locally, confirm extracted values manually, and review changes without treating them as proof of fraud.
The Invoice Fraud Change Inspector compares two versions of an invoice across 13 extracted fields — invoice number, date, vendor, tax ID/GSTIN/VAT, bank account, IBAN, UPI ID, currency, subtotal, tax amount, total, line-item count and line summary — and marks each one added, changed, removed, unchanged or unavailable. Anything in the payment-routing group (bank account, IBAN, UPI ID) is flagged separately, because that is where a redirected payment shows up. It is for accounts-payable staff and finance teams doing a manual second look, and it deliberately reports observable differences only, never a fraud verdict.
Open Invoice Fraud Change Inspector on AltFTool — it loads instantly in your browser.
Paste invoice text or a JSON invoice object into "Earlier or trusted invoice" and "New invoice to inspect", or press Choose text PDF to read a local PDF's text layer, then press Extract fields on each side.
Correct every value under Confirm earlier invoice fields and Confirm new invoice fields, tick "I reviewed these extracted fields against the original invoice" on both, and press Compare confirmed fields.
Metric cards report Fields reviewed, Observable changes, Routing changes, Unchanged and Unavailable; a changed bank account, IBAN or UPI ID raises a "Payment-routing details changed" panel, and Download change counts saves altftool-invoice-change-counts.json with counts only.
Bank account, IBAN and UPI ID sit in their own group, so a changed beneficiary is never buried among harmless edits like a reworded line summary.
Amounts are compared numerically rather than as strings, and account numbers, IBANs, tax IDs and UPI IDs are normalised for spacing and case so cosmetic reformatting is not reported as a change.
You must tick that you have checked the extracted fields against the original document on both sides before a comparison runs, so a bad PDF text layer cannot quietly produce a wrong diff.
No. It reports observable field differences between two invoice versions and flags which of them touch payment routing; a change is not proof of fraud, intent or identity. Treat the output as a prompt to verify the account with the supplier through a known phone number, and escalate through your organisation's own controls.
Thirteen, split into four groups: identity (invoice number, invoice date, vendor, tax ID/GSTIN/VAT), payment routing (bank account, IBAN, UPI ID), amounts (currency, subtotal, tax amount, total) and line items (line-item count, line summary).
It can read the text layer of a text-based PDF locally in your browser, and it also accepts pasted invoice text or a JSON invoice object. A scanned or image-only PDF has no text layer, so nothing will be extracted and you will need to type the fields in and confirm them yourself.
No. Extraction and comparison run in the page, and the exported report is a counts-only JSON file that records how many fields in each group were added, changed, removed, unchanged or unavailable — it does not include the invoice values themselves.