Check a Pay Stub, Statement, or PDF for Signs of Editing

Drop a document, tell us what it's supposed to be, and get a ranked list of risk indicators with the evidence behind each one — deterministic forensic checks, not an AI guess, entirely on your device. Nothing is uploaded.

How It Works

1

Drop a Document

A PDF or a photo of a pay stub, bank statement, invoice, receipt, or utility bill. It never leaves your tab.

2

Say What It Claims to Be

Pay stub, bank statement, invoice, or other — this picks the right math checks and sets expectations for who should have produced it.

3

Local Forensic Checks Run

Structure, software fingerprints, fonts, hidden text, math, dates, and image signals are checked in seconds — no model download, nothing sent anywhere.

4

Ranked Findings With Evidence

A risk score and band, every finding explained in plain language with clickable evidence, the checks that passed, and a printable report.

What It Checks

Document Structure & Earlier Versions

Many edited PDFs still carry their own earlier versions inside the file. When one is found, it's rebuilt and compared side by side with the current version, so you see exactly what changed — not just that something did.

The Software That Made It

Every PDF names the software that produced it. That name is checked against known payroll/statement systems, ordinary photo and PDF editors, and known paystub-generator services — and the edit history is checked for editing software appearing after the document was supposedly finalized.

Fonts & Hidden Text

Numbers set in a font that appears nowhere else on the page, text pasted onto a scanned image, and text hidden underneath a covering box are all detected — and any hidden text is recovered and shown to you.

The Math

Gross minus deductions should equal net, to the cent. Tax withholding should sit in an expected range. Year-to-date totals should track the pay period. Routing numbers, IBANs, and card numbers are checked against their own official checksums, not just their shape.

Dates

Creation and last-modified dates are compared to each other and to the dates the document itself claims — and to transaction dates inside the content, so an out-of-order or impossible date doesn't slip through.

Image & AI-Generation Signals

Photos are checked for editing-software fingerprints and re-compression signs. Images also get checked for a signed "Made with AI" label — Content Credentials — and for generation details some AI tools leave behind, sometimes including the exact prompt used to create the image.

Privacy Guarantee

  • Every check runs locally in your browser — your document is never uploaded
  • Nothing about your document is stored anywhere — close the tab and it's gone
  • The disconnect test: load the page, then turn off your Wi-Fi — screening a document keeps working with zero connection

Who This Is For

  • Landlords & property managers: a second look at an applicant's pay stub or bank statement before enterprise screening tools are worth the cost
  • Bookkeepers & small-business owners: check an invoice's bank details or a receipt before it's reimbursed
  • HR & small lenders: a first-pass income-verification read without a per-document API contract
  • Marketplace buyers & sellers: check a proof-of-payment screenshot or PDF before you ship or release funds

How Scoring Works

The score is a plain weighted sum, not a learned model — every constant below is the exact one the tool uses, and every finding shows its exact point contribution. We publish it because a screening tool that shows its work is one you can actually trust.

Points per finding, by severity

  • Critical — hard evidence of tampering60 pts
  • Strong — unlikely on a genuine document25 pts
  • Moderate — consistent with editing, but common innocently8 pts
  • Info — context only0 pts

Per-category point caps

  • Document structure70
  • The software that made it45
  • Fonts & hidden text60
  • The math50
  • Dates16
  • Image signals25
  • AI-generation signals60

Category caps exist so a handful of weak signals in one area can never add up to look like strong evidence.

0–19
Low
20–54
Elevated
55–100
High
  • · Any single critical finding forces the band to High, no matter what the total adds up to.
  • · Image-based findings alone can never push the band past Elevated — image evidence is the least reliable signal here, and the score reflects that humility.

A low score does not authenticate a document. A document made from scratch with a paystub-generator site can be structurally clean and still not be genuine.

A high score is not proof of fraud. Always verify with the document's issuer before acting on it.

What This Can't Do

  • · It never labels a document "fake," "forged," "authentic," or "genuine" — only risk indicators, with evidence and possible innocent explanations. You make the final call.
  • · A low score doesn't authenticate a document, and a high score isn't proof of fraud — both need human judgment and, where it matters, verification with the issuer.
  • · Scanning or photographing a document destroys most of the digital evidence a screening depends on. Ask for the original file when you can.
  • · It does not screen identity documents — passports, driver's licenses, or ID cards. Financial documents only.
  • · It can't tell you an image is AI-generated once its disclosure markers have been stripped — which is common and easy to do. Absence of a disclosure proves nothing either way.
  • · It never tells you how to make a document score lower. There is no remediation guidance, ever.

Frequently Asked Questions

How can you tell if a pay stub is fake? +
We don't declare a pay stub fake — we run a set of deterministic checks and show you what they found. That includes whether the gross-to-net math actually adds up, whether tax withholding is in the expected range, whether the document was produced by a paystub-generator service or a photo editor rather than a payroll system, whether the file contains an earlier version of itself with different numbers, and whether any text was hidden underneath a covering box. Each check shows its evidence and a possible innocent explanation — you make the call.
Can you detect if a PDF was edited? +
Often, yes. Many edited PDFs still contain their own earlier versions inside the file — when that's the case, we recover them and show you a before/after comparison of exactly what changed, which is much stronger evidence than a single metadata field. We also check the software listed as having produced the file and look for numbers or text that were added later. A document that shows none of these signs isn't proven untouched — see "Is a high score proof of fraud?" below.
Can you tell if an image is AI-generated? +
When the tool that made the image disclosed it — a signed Content Credentials label, or generation details left behind by the generator itself — we detect that disclosure and show it to you, sometimes including the exact instructions used to generate the image. When that information was stripped, which is common and easy to do, no tool can reliably tell you whether an image is AI-generated, and we say so instead of guessing. We do not run a pixel-based "AI or not" classifier — those are unreliable and we won't dress up a guess as a finding.
Is a high score proof of fraud? +
No. A high score means several indicators consistent with editing were found — it's a reason to look closer and verify with the document's issuer, not a verdict. Likewise, a low score does not authenticate a document: a fake built from scratch with a paystub-generator site can pass every structural check and still not be genuine. The score exists to help you triage a pile of documents, not to be quoted as proof.
Is this legal to use for tenant screening? +
This tool reports indicators about a document, never a judgment about a person, and screening indicators should be applied the same way to every applicant and confirmed with the document's issuer before any adverse decision. We're not a law firm and this isn't legal advice — check your local tenant-screening and fair-housing rules before building it into a process.
Why didn't my scanned document show much? +
Scanning or photographing a document destroys most of the evidence a screening depends on — the earlier-version history, the original producing software, the original text layer. A print-then-scan is also a classic way to launder an edit. If you have the option, ask for the original digital file (PDF or the bank's own export) rather than a scan or photo of a printout.