Upload a bank statement as a PDF or a photo of one. PDF Insights pulls out the transactions, checks the balances against each other, and tells you what doesn't add up — before you underwrite it.
Drop in a PDF or a photo of a statement. There's no template to configure and no bank to pick from a list first.
Dates, descriptions, amounts, and running balances are pulled line by line and checked against each other for arithmetic that doesn't reconcile.
A 1–100 score comes out the other side, with every flag that fed into it listed alongside — not just a number to trust blindly.
The score isn't a black box. It's built from a short list of concrete checks run against the actual line items on the statement.
Each transaction's amount is checked against the change in balance from the line before it. Statements where the numbers don't chain together get flagged rather than trusted.
Transfers between accounts, loan proceeds, and other non-sales deposits are filtered out, so the revenue figure reflects money actually earned by the business.
Every day the account sat below zero, or below a thin cushion, is counted — a better read on cash-flow stress than a single point-in-time balance.
The same date, description, and amount showing up more than once is a common sign of an edited or spliced document, and gets called out by line.
Upload a bank statement PDF, or a clear photo/scan of one. Everything runs in your browser — nothing is uploaded to a server.
PDF, PNG, or JPG
| Date | Description | Amount | Balance |
|---|
This demo parses the text layer of your document (or runs OCR on images) and applies a general-purpose line parser plus the checks described on the overview page. Statement layouts vary widely — treat results here as illustrative of the workflow, not as a production-grade underwriting engine.