Image and document fraud detection for lending
A small-business loan file is mostly pictures: the shop front, the stock, the invoices, the bank statement, the ID and the selfie. TrueSecurix checks each one and tells your credit team whether to pass it, review it or flag it, with the evidence behind the answer.
In our held-out test of 31,966 images, 0 of 2,526 genuine lending photos (shops, markets, receipts, documents) were flagged and 87.8% of manipulated images were caught. Read how we measured
What credit and risk teams at lenders and NBFCs are seeing in loan files now.
What fraud looks like in a loan file today
Shops that do not exist
A generated kirana store, or a real shop with its signboard renamed to match the application, stands in for a business nobody has visited.
Stock that was never bought
Empty shelves filled in by an AI editor, or a warehouse photo made from a prompt, to support a working-capital limit.
Invoices and statements with new numbers
An amount changed on an invoice or a receipt, or a balance changed on a bank statement, to inflate turnover or income.
A borrowed face
A selfie with someone else's face swapped in, or a person who does not exist, paired with a stolen or made-up ID.
Image and identity checks run on the same account and the same credits: one credit per check.
Which checks apply, and what you get back
Each image goes to POST /v1/image/authenticity; selfies, IDs and statements go to POST /v1/verify. Every response carries an outcome and the evidence behind it, and every check is written to your history.
| What you receive | Check | What the response gives |
|---|---|---|
| Shop and stock photos | Image authenticity | Real, AI generated, AI edited or digitally edited, with the indicators and the located edited areas. |
| Invoices and receipts, as photos or scans | Image authenticity | The same four answers, with the edited areas located so a reviewer knows where to look. |
| Bank statements and salary slips, as text-based PDFs | Bank statement reconciliation | The row where the running balance stops adding up, or the payslip total that does not reconcile. |
| Borrower selfie and ID photo | Identity checks | Whether the selfie is face-swapped or generated, and whether it matches the photo on the ID. |
| Aadhaar, PAN and passport | Checksum validation | The field and the reason when a number fails its own check digits or format. |
Nobody is rejected on a guess. The model tells you which files a person should look at.
Three outcomes your credit team can act on
The decision field turns the answer into an action. Route on it directly in your loan origination system.
Accept it
No sign of generation or editing. The file continues through your normal credit checks.
A person decides
Some evidence points to a manipulation, not enough to accuse a genuine borrower. A credit officer takes a short look, guided by the class it leans toward.
Treat as manipulated
Strong evidence of generation or editing, with the indicators and the edited areas. Hold the file and ask for the original or a visit.
Context, not a compliance claim.
Where this sits in your due diligence
Under the RBI's digital lending rules, the regulated entity remains responsible for assessing the borrower, including when a lending service provider runs part of the journey. TrueSecurix is one evidence check inside that assessment. It does not replace your due diligence, your KYC obligations or your credit policy, and it never makes the credit decision.
How TrueSecurix fits the rules a regulated business works under
How a team rolls it out
Most of the work is wiring three outcomes into the loan origination system you already run. You can build all of it before you spend a credit.
One request, one image, one answer
curl https://truesecurix.com/v1/image/authenticity \ -H "X-API-Key: $TRUESECURIX_API_KEY" \ -F "image=@shop_front.jpg" \ -F "reference=loan-48211"
Field by field in the API reference.
- Build against a test key. A test key returns a simulated answer at no charge, including the exact review shape when you send
reference=review, so your loan origination system handles all three outcomes before any live image is sent. - Run a proof of concept on your own labelled images. A folder of real and manipulated shop photos, invoices and statements from your own files, with a CSV of labels, goes through the batch script. You get accuracy per class, the confusion matrix, genuine images wrongly flagged and the share sent to review, on your data.
- Send results into your loan origination system. Signed webhooks post
check.flagged,check.reviewandcheck.completedto your HTTPS endpoint, signed with HMAC-SHA256 and retried for about a day. - Route reviews to a person. A review carries the class the evidence leans toward and the indicators behind it, so a reviewer knows what to look for. Every check stays in your history with its model version, and exports to CSV.
Measured on images it never saw
The image model was measured on business photos of shops, markets, receipts and documents, plus general photos, in a held-out test of 31,966 images never used in training, threshold setting or model choice.
Your own images are the measure that counts for your use. A proof of concept on your own labelled images gives you the same figures on your data, with the per-image CSV behind every number.
Questions risk, operations and engineering teams ask first.
Questions, answered
Does TrueSecurix make the credit decision?
No. Each image or document comes back as pass, review or flag with the evidence. Your credit policy decides what happens next, and a review always goes to a person.
Which loan documents can it check?
Any JPEG, PNG or WebP image up to 15 MB: shop and stock photos, invoices, receipts and document photos. Bank statements and salary slips are sent as text-based PDFs to the identity endpoint, which re-adds their totals.
How accurate is it on our own loan files?
In our held-out test of 31,966 images, 0 of 2,526 genuine lending photos (shops, markets, receipts, documents) were flagged and 87.8% of manipulated images were caught. A proof of concept on your own labelled images gives you the same figures on your data.
Do you keep borrower images?
No. Images are analysed in memory and never stored. We keep the SHA-256 fingerprint and the result, so you can prove which file produced which answer. The service is hosted in India.
Check the images in your loan files
Try a photo in the live demo now, or create an account for 100 free checks. No card is needed, and test keys are free while you build.
Running lakhs of checks a month? See enterprise terms and a proof of concept.