Fund the loan their account supports
A payment size and the stacked loans a bureau pull misses, returned while the applicant is still on the page.
Instant result before your applicant shops elsewhere.
One click to connect a bank, or a statement upload. That is everything your applicant has to provide, and 90% of the time they get their result in ten seconds.
A payment size, not just a risk level.
KoraConnect tells you how much this applicant can afford, based on the policy you set.
See how the thresholds are setOne number, off verified income and the obligations we find in the account, at the cap you set.
Every threshold is a setting, not a model, and each one that trips comes back with the value that crossed it.
Stop the loan you would have stacked.
KoraConnect finds the payments a bureau file never shows you, so you are not lending on top of debt nobody has filed yet.
Buy-now-pay-later plans, cash advance apps, earned wage access and rent are all repayments, and none of them reach a bureau file.
Every obligation carries its status, so a loan that stopped being paid does not sit in the ratio as if it were live.
The money landing is read as well as the money leaving, so a new loan shows up before its first payment does.
Approve on income a payroll file cannot show.
KoraConnect finds the income a payroll file misses, so you can approve applicants you would otherwise have to decline.
Verify every applicant, not just the salaried ones
Delivery platforms, freelance work paid by transfer, benefits, private support. A payroll database reaches under half of Americans with payroll income; between a bank connection and statements, this reaches every applicant.
Stop counting borrowed money as income
Summing deposits counts advances and transfers as earnings, and it overstates worst for the applicants you most need to price correctly. Loan disbursements, movement between the applicant’s own accounts and large one-off deposits come out before the figure is formed.
See how the categorization worksKnow the file is real before the money leaves.
KoraConnect reads the document and the account behind it, so a doctored statement, a total that does not add up, or reversals and cash-advance draws in the account come back in the same result as the income.
See how document fraud detection worksA risk score that moves when the bureau file does not.
KoraScore is read off the applicant’s own account in real time, so it carries signal the bureau file does not and updates when the account does.
Read in real time off this month’s transactions rather than what lenders have filed, with the drivers named, on the familiar 300 to 850 scale.
KoraScore is built from the account rather than from filed tradelines, so inside a single bureau band it still tells your safest applicants from your riskiest.
Choose the outcome it predicts, charge-off or 60 days past due, and take it as one input to the model you already run. It never returns a decision, and the cutoff stays yours.
Answer them inside the application they started.
One result, three ways out: into your decisioning the second it finishes, onto a dashboard when someone wants to look, and onto a page for the file.
The webhook fires the moment the analysis finishes and one call returns the figures, which is what makes a ten-second answer reach the applicant rather than an inbox the next day.
Every transaction with the label Kora put on it, so an underwriter can see which deposits counted as income and which debits were read as debt.
A dated document holding the figures the decision ran on, so the file still explains itself long after the session closed.
Credit desk questions.
The six that come up before a pilot.
01How fast does a cash flow analysis come back?
Under 45 seconds end to end, and 90% finish in under 10 seconds. That’s fast enough to give someone an answer while they’re still in the session, instead of emailing them a decision the next day.
02Do we keep our own credit policy?
Yes. The payment-to-income cap, how income’s calculated, which obligations count, every threshold, and which risk indicators are even turned on: all of that gets configured per lender, then applied the same way to every file. We compute the condition and tell you what tripped it; where you draw the line is entirely up to you.
03What if an applicant will not connect a bank account?
They can upload statements or paystubs instead. We don’t just take those at face value: the documents get extracted, the arithmetic gets reconciled, and we check for signs of editing. You can also send us raw transaction data from an aggregator you’re already using. Between a live connection and statements, we’ve got every applicant covered.
04What happens with a no-hit or thin-file applicant?
Nothing about the process changes. A bureau needs a repayment record to say anything at all; cash flow just needs a bank account. You get the same income streams, obligations, residual, and score on a file with no credit history as you would on a thick one.
05How does a finished analysis reach the system our underwriters work in?
A REST API with signed webhooks, a dashboard with the full labeled transaction ledger, or a PDF for the file. Most lenders pull in the verified income, the obligations, the risk indicators and the score straight into a scorecard they already run.
06How do we validate the score against our own book?
We backtest it on loans you have already decided. You send historical originations with how they performed and your own definition of a bad loan, and we report how the score ranks that book, which signals drove it, and where the cutoffs would have fallen. You set the cutoff from your own results rather than ours.
See the payment that holds, on a live application.
Book a demo and we’ll take one application end to end: the verified income, the debt no report filed, the payment it supports, and the result your customer sees in ten seconds.