Login
Book a Demo

Patented Bank Statement Fraud Detection

Catch altered bank statements, even after the metadata is gone

A fraudster edits a few deposits, re-saves the PDF, and metadata-based tools see a clean file. Thumbprint examines the
document itself and scores its authenticity from 1 to 1,000 in under five seconds.

Book a demo

Two models, one score

Both models feed one score, with the reasons listed. Send PDFs by API or drag and drop, and get results in the platform or by webhook.
ThumbPrint Statement Reconciles Risk Page

Does the statement agree with itself?

Checks fonts, alignment, spacing, date formats, and whether the balances reconcile.

Does it match the bank's real statements?

Compares the file against verified statements from the same bank to catch forgeries that look consistent but aren’t built the way that bank builds them.
Read the API docs
~ 0 M+
statements scored
~ 0 M+
altered documents flagged (past 2 yrs)
~ 0 M+
first-time bank statements per month
Protected by U.S. Patent 12,260,174 B2, “Detection of Altered Documents.”

From Upload to Score

Upload Documents Icon

Upload

Send a PDF statement through the API or drag and drop it in the web platform.

Complete Application Icon

Analyze

Both models run at the same time.

Read the API docs
Review Scorecard Icon

Score

Get the authenticity score and contributing factors in seconds, in the platform or by webhook.

A score you can act on

ThumbPrint Fraud Detection
Review a flag in seconds. Set score thresholds to determine what’s next.

Thumbprint vs. Metadata-Only Tools

Fewer false alarms and fewer missed forgeries, so you fund on statements you can trust. underwriters spend their time on the files that deserve it.

Metadata-only tools

Thumbprint

Works after a file is re-saved or scrubbed
Examines the document’s content and layout
Compares against real statements from the same bank
Output

Pass or fail flag

Score from 1 to 1000 with the contributing factors

Reference library grows as new statement formats are verified

Key Capabilities

Authenticity score from 1 to 1000

with ranked contributing factors

Results in seconds

typically under five

Works with statements

from thousands of banks

API and webhooks

to send scores straight to your LOS, CRM, or decision engine

Configurable thresholds

that fit your review and decline rules

Runs alongside PDF Insights

so every statement is converted and verified in one pass

Built for Lenders Who Fund on Statements

Industry

How they use Thumbprint

Link

MCA funders
Screen every submission before underwriting
MCA
Small-business lenders
Verify statements inside the credit workflow
Small Business Lenders

FAQ

Thumbprint is MoneyThumb's patented fraud detection for bank statements. It scores each PDF statement from 1 to 1,000 for authenticity and lists the factors behind the score.

Yes. Thumbprint examines the document's content, layout, and structure rather than relying on file metadata, so re-saving or scrubbing a file doesn't hide the alteration.

Two models run on each statement. One checks whether the document is internally consistent. The other compares it against verified statements from the same bank. Both feed one weighted score.

Most statements are scored in under five seconds.

Yes. Send statements and receive scores through MoneyThumb's REST API and webhooks, or use the web platform. See the docs at docs.moneythumb.com.

No. MoneyThumb never trains models on customer data. MoneyThumb is SOC 2 Type II compliant and hosted in the US.

Security

SOC 2 Type II compliant · Hosted in the US on AWS · Customer data never used to train
Learn More

Fund on statements you can trust

Most fraud tools check file metadata, which disappears the moment a PDF is re-saved or scrubbed. Thumbprint examines the document itself and gives every statement an authenticity score from 1 to 1000, usually in under five seconds.

Book a demo
Skip to content