Resistant AI checks whether a document was forged. Most lenders also need to know whether the numbers on a real statement support the loan. Upload a statement below and see both checks, from $99 a month.
Upload a document to extract
Drop files here or click to upload
Up to 50 files
Uploading...
Upload a bank statement and watch the analysis run live, free, no signup required.
Resistant AI sells document forensics. Its Resistant Documents product looks at how a PDF or image was built, not what it says, and returns a verdict of Trusted, Warning or High Risk. The company says each document is checked more than 500 ways, and it is good at the fraud a human reviewer cannot see: edited fields, reused templates, and statements produced by image generators. Enterprise risk teams at payment companies, marketplaces and European lenders run it at high volume.
The price is the reason most US lenders start looking at alternatives. Resistant AI publishes no price on its own site (resistant.ai/pricing returns a 404), but its AWS Marketplace listing for Document Forensics shows $10,000 a month for each of four analysis types on a one-month contract, or $120,000 for a 12-month contract on one type. Custom quotes go through a private offer. For a lender reviewing a few hundred files a month, that is a large fixed cost for one check in the underwriting file.
The second reason is scope. A forensic verdict tells you a statement is authentic. It does not tell you the deposits are inflated by transfers from the owner's other account, that the borrower is paying three merchant cash advances, or that the average daily balance cannot carry the new payment. A genuine statement can still tell a false story. The alternatives below split along that line: forensic engines that answer is it real, and analysis tools that answer does it support the loan.
Every fact about Resistant AI on this page comes from its own website, its AWS Marketplace listing or a named directory, so you can check each one before a vendor call.
Resistant Documents is language and template agnostic, so it does not need a bank's layout on file to flag an edit. It looks for the traces online editors leave, for templates bought from fraud farms and reused across applicants, and for the textures image generators leave in AI-made statements. Verdicts come with the evidence behind them rather than a bare score, and the company can add device and behavior signals, which it says catch up to 30% more fraud than the documents alone. Its site also sells Resistant Transactions, an AML and payment fraud product with more than 80 prebuilt models.
The Document Forensics listing (sold by Resistant AI on AWS) prices four dimensions separately: Document Forgery Analysis, Customized Forgery Analysis, Document quality analysis and Document classification analysis. Each shows $10,000 per month on a one-month contract, with 12, 24 and 36 month terms at $120,000, $240,000 and $360,000 per dimension. The listing adds that fees paid are non-refundable and invites buyers to request a private offer for custom pricing. Treat these as list prices a negotiation starts from, not what every customer pays.
The products page on resistant.ai says verdicts can be customized to your risk appetite with 99.2% accuracy. The AWS listing, written by the same company, says approvals and declines are automated with 99.92% accuracy. The ROI calculator FAQ uses different numbers again: an 86% day-one fraud detection rate before any training and a 92% average reduction in manual reviews. None of these is wrong on its own terms, but they measure different things. Ask which figure applies to US bank statements in a pilot on your own files.
Capterra lists Resistant AI with zero reviews, contact-vendor pricing and no free trial, and suggests Persona, Veriff, Plaid and IDnow as alternatives, which are identity verification tools rather than document forensics. The AWS listing carries one review at 4 stars. PeerSpot has one review at 4.0 from a founder who called the pricing relatively expensive for their market while noting it was acceptable for a large bank. For a buyer, that means there is little public evidence either way, so a paid pilot on your own document mix matters more than usual.
Most loan fraud on bank statements is not a forged PDF. It is a real statement from a real bank where the borrower moved money between accounts to inflate deposits, took an advance last month and left it off the application, or timed deposits to lift the balance on the days that get printed. Every one of those files passes a forensic check as Trusted. Catching them takes transaction-level analysis: separating revenue from transfers, grouping recurring debits by counterparty, and reconciling every line to the printed balances.
LenderAnalyzer reads each statement to the transaction, reconciles it to the opening and closing balances, and flags any month that does not tie out, which is how an edited amount or a missing page shows up. It then computes true revenue net of transfers and loan proceeds, average daily balance from every day, NSF and negative days, and existing debt grouped by lender or funder. Every number links back to the transactions behind it, and the output goes to Excel or the API.
If you onboard thousands of documents a month from anonymous applicants, face organized template fraud, or need one fraud engine across KYB, claims and lending in several countries, a dedicated forensic vendor is the right buy and Resistant AI, Inscribe and Fortiro belong on the shortlist. Many lenders run a forensic check and a cash-flow analysis side by side. The decision is whether the forensic layer is worth a five or six figure contract at your volume, or whether reconciliation and transaction analysis already catch the fraud you actually see.
Prices are each vendor's own published figures, from its site or its marketplace listing. Where a vendor publishes nothing, the table says so.
Swipe sideways to see the full comparison
| Tool | What it checks | Published price | What it leaves to you | Best for |
|---|---|---|---|---|
| LenderAnalyzer This page | Statement math reconciled to printed balances, true revenue, ADB, NSFs, existing debt and stacked advances, with every figure traceable | From $99 a month, with volume and enterprise tiers | Pixel and metadata forensics on the file itself | US lenders and funders who need the underwriting numbers and the math-level fraud check in one upload |
| Resistant AI | Document forensics (forgery, templates, AI-generated files), document quality and classification | Nothing on its site; AWS Marketplace lists $10,000 a month per analysis type, $120,000 for 12 months | Revenue, balances, debt and every other underwriting number | High-volume risk teams fighting organized document fraud across several markets |
| Inscribe | Forensic, network and AI-generation checks on bank statements, pay stubs and tax forms, plus income verification | Not published; quote based on document volume | Spreading and cash flow analysis | US fintech and bank risk teams whose main job is document authenticity |
| Ocrolus | Lending-grade extraction with cash flow and income analytics and a fraud signal layer | Not published; metered per document or per file on a sales quote | Price transparency for small teams | Larger lenders standardizing on one extraction and analytics vendor |
| HTPBE | PDF tamper detection only, one check per PDF from a web app or API | $15 a month for 30 checks, $149 for 350, $499 for 1,500 | Everything beyond whether the PDF was edited | Small teams that want a cheap self-serve tamper check |
| Fortiro | Fake and manipulated income and identity documents, with template and metadata checks | Not published on its pricing page | Transaction analysis and underwriting metrics | Lenders whose fraud is concentrated in pay stubs and income documents |
| Manual review | What an underwriter notices by eye, such as fonts, alignment and totals | Analyst time; Resistant AI itself puts a manual document check at about $2 to tens of dollars | Edits invisible to the eye and math across hundreds of lines | Very low volume with experienced reviewers |
Comparison compiled by LenderAnalyzer from public vendor materials; see the date noted above each table. Competitor names are trademarks of their respective owners; figures may change, so verify current details with each vendor.
Computed deterministically from every extracted transaction, every figure traceable to its source line.
Computed across the full statement period, carried forward day by day.
Deposits vs withdrawals and net flow, broken down month by month.
Every insufficient-funds and overdraft incident counted, with fees totaled.
Recurring deposits grouped into income streams with estimated monthly amounts.
Debits to other lenders and funders detected and totaled per month.
Days below zero across the period, a direct stress signal.
The biggest credits with dates and sources, concentration flagged.
Automatic red and yellow flags your analysts can review in seconds.
Drop in PDFs, scans or photos, one statement or a multi-month package, from any bank.
Every transaction is extracted, then cash flow, balances, income streams, NSF activity and debt payments are computed.
Read the underwriting snapshot, download the Excel report, or pull structured JSON into your LOS via API.
28 lending document types extracted out of the box, build the complete picture of an applicant's financial situation.
Common questions from lending and credit teams.
Resistant AI is a fraud detection company that sells two products. Resistant Documents checks bank statements, pay stubs, tax forms, invoices and other files for forgery, template reuse and AI generation, and returns a verdict of Trusted, Warning or High Risk. Resistant Transactions adds AI models to an existing transaction monitoring system for AML and payment fraud. Its buyers are mainly payment companies, marketplaces, insurers and lenders.
Resistant AI does not publish pricing on its website, and resistant.ai/pricing returns a 404. Its AWS Marketplace listing for Document Forensics shows $10,000 per month for each of four analysis types on a one-month contract, and $120,000 for a 12-month contract on one type, with custom quotes through a private offer. Fees are listed as non-refundable, so confirm scope and volume before you sign.
The closest competitors in document fraud are Inscribe, Fortiro, Ocrolus with its fraud layer, and smaller self-serve tamper checkers such as HTPBE. In transaction monitoring it competes with firms like Featurespace and Hawk. Capterra's alternatives list shows identity verification tools such as Persona and Veriff, which solve a different problem. For lenders, the useful comparison also includes cash-flow analysis tools like LenderAnalyzer, which catch the fraud a forensic check passes.
It depends on the fraud you see. If your losses come from forged or AI-generated PDFs at high volume, Inscribe or Fortiro are the closest like-for-like options. If they come from real statements with inflated deposits, hidden advances or balances that cannot carry the payment, a transaction-level analysis tool such as LenderAnalyzer catches more of it and costs from $99 a month.
Both are document fraud engines for financial services, and neither publishes a price on its website. Resistant AI stresses that it reads how a document is built rather than its content and sells a separate AML product. Inscribe focuses on US fintechs and banks, adds a network check across documents seen at other institutions, and offers income verification. Run both on the same sample of your files before choosing.
Yes. Resistant AI sells Document Forensics through a REST API, and its AWS listing says it can be deployed with 14 lines of code. It is also listed on Google Cloud Marketplace and Salesforce AppExchange. The API returns the verdict and the evidence, so your system or a reviewer still decides what to do with a Warning file.
Resistant AI says its model detects textures, structural patterns and other traces left by the major image generators, and it markets a dedicated Gen AI detection use case. That is its strongest ground. An AI-made statement built with correct arithmetic and a plausible transaction history is harder, which is why lenders pair any forensic verdict with a reconciliation and cash-flow check.
The fastest reliable test is arithmetic: add every deposit and withdrawal to the opening balance and compare the result with the printed closing balance for each month. A statement that does not tie out was altered or is missing pages. After that, lenders compare deposits with tax-reported income, strip out transfers between the borrower's own accounts, and look for recurring debits the application did not disclose.
No. A forensic tool answers whether the document is authentic. The approval still depends on whether revenue, balances and existing debt support the payment, and most loan fraud sits in real documents that tell a misleading story. Treat the forensic verdict as one input and the cash-flow analysis as the underwriting decision.
Yes. The tool at the top of this page runs on a real statement, so you can see the reconciled transactions and the summary before creating an account. Paid plans start at $99 a month, with volume and enterprise tiers, and there is no annual contract required to start.
How credit teams run these calculations by hand, so you can see exactly what the software automates.
The other quote-only fraud engine, compared for lenders.
Math, transfer and hidden-debt checks on every statement.
What drives an Inscribe quote and how to judge it.
The checks an underwriter can run on any file.
Extraction plus analytics without a sales-gated price.
Another enterprise vendor priced through AWS Marketplace.
Undisclosed advances grouped by funder.
Per-page and per-document rates side by side.
Income documents checked against the deposits that pay them.
Analyze your first statements free, plans from $99/month, 50% off billed annually.
