Fake receipts: how to spot them
There are websites where you type a shop, an amount and a date and get a receipt that looks real. Here is how to tell the difference, by eye and inside the file.
Updated 2026-08-31 · Veriflied
How fake receipts are made today
There are two kinds. Template generators: websites with ready-made designs for known retailers, where you fill in the fields and download the result as a PDF or an image. And AI image models, which produce a 'photo' of a crumpled receipt on a table, complete with shadows and slightly blurred print. Both are used in insurance claims, expense reimbursement and webshop returns.
The two leave different traces, and it matters which one you are looking at before you assess it.
Signs you can see yourself
Receipt content
Real receipts carry a till number, cashier number, transaction id and often a receipt number that can be looked up in the retailer's system. Generators fill the fields with random but tidy numbers.
The retailer's format
Every chain has its own receipt layout, and it changes over time. A receipt dated 2024 in the 2021 layout is a red flag. Compare with a real receipt from the same period.
Numbers and tax
Do the maths. Tax, discounts and totals must agree. Generators rarely calculate correctly, and AI models do not calculate at all.
The photo
A photo of a receipt needs a background that makes sense, matching perspective, and paper fibre and print noise that look like real thermal paper. AI images often have perfect paper and slightly warped text at the edges.
The story
Does the receipt fit the purchase date, payment method and the rest of the case? A card purchase can be confirmed against a bank statement. A large 'cash purchase' is unusual in itself.
What the file reveals
Template receipts downloaded as PDF were made by a browser or an HTML-to-PDF tool and carry no issuer identity. They are typically created minutes before submission. Timestamps, producer chain and missing metadata together say the document was made by a person, not by a till.
Downloaded as an image, the PDF traces are gone, but the file does not look like a real photo either: no camera information, no capture time, and a format and size that match a screen, not a phone.
AI-made receipt photos are caught by the image analysis itself. An AI detector assesses whether the pixels were generated and a heatmap shows where. Combined with metadata and timeline, that gives an overall assessment with reasons.
Good habits if you handle many receipts
Always ask for the original file, not a screenshot. Keep the file without opening and re-saving it. Cross-check large amounts against a bank statement. And run documents above a certain amount through a forensic analysis before approval. It takes seconds and catches what eyes cannot.
Frequently asked questions
Are receipts from generator sites illegal?
The sites often market themselves as 'fun' or 'for testing'. Using a fabricated receipt to obtain money is fraud.
Can you tell from a photo that a receipt is AI-generated?
Yes, an AI detector is built for exactly that and finds both fully generated images and inserted regions. It looks at pixel patterns, not at whether the receipt looks right.
What if the receipt was sent via a messaging app or email?
Then some metadata was stripped on the way, which is normal. The analysis accounts for it and relies on the traces that survive: producer chain, PDF timestamps and the image content itself.
Can I try it on a receipt I am unsure about?
Yes. The first 10 analyses in VeriFly are free, and the report shows the concrete findings.
Try it on your own document
Create an account and analyse your first 10 images or documents for free. You get a report with the concrete findings, not just a score. All processing in the EU.