Plans from $99/month. True revenue, positions, and 27 fraud signals included.
27 signals across two layers on every file: forensics on the PDF itself, then ClearStaqAI's visual check against real bank templates. Every finding is explained in plain English.
| Severity | Factor | Weight |
|---|---|---|
| Suspicious | Different fonts in same column | 6 |
| Suspicious | Different font sizes in same column | 6 |
| Suspicious | Inconsistent text operators | 4 |
fraud signals checked on every statement, across two layers.
fraud score per document, banded Low, Medium or High.
extra per file when ClearStaqAI's visual check is switched on.
How it works
Before reading a transaction, ClearStaq inspects the PDF: who produced it, when it was modified, and whether fonts and figures hold together.
The page is rendered and compared with authentic templates for that bank, catching pixel-level edits that leave no trace in the file.
Each finding carries a severity and a sentence that says why. Findings roll up across every statement the merchant sent.
A producer string from an editing tool means someone touched the file after the bank made it.
Banks generate statements once. A file modified days after it was created is a red flag, with both timestamps shown.
Fonts and sizes that change inside one column often mean a number was typed over the original.
Across the whole application
Findings from every statement a merchant sends roll up into one risk profile, next to true revenue, NSFs and negative days.
What to do with a flag
FAQ
Yes. The forensic layer runs on every statement on every plan.
A visual comparison against authentic bank templates, which catches edits that leave no trace in the PDF's structure. Switch it on per upload; it adds about 10 seconds per file.
Scans are flagged as scans, because the file itself says less. The visual check matters more there.
It means it needs a human look. Every finding says what was found so your team can decide.