EFBacked by Entrepreneur First

Underwrite borrowers anywhere in the world, in minutes.

LendTrace reads borrower documents, completes incomplete files, and drafts cited credit memos — so your officers can focus on the call that matters.

Live in PH · ID · MX · ZA · US
Vision-language extraction
1.4s
GCashPHP
Available balance
₱ 38,420.50
+₱ 1,200 this week
Photo · at angle
Structured fields · 12/12 ✓
  • account_holderMaria Santos
  • available_balance38420.50 PHP
  • monthly_inflow52,310.00
  • tamperingnone · 0.04
✓ Account match✓ Balance reconciled⚑ 1 cross-doc flag
100K+
borrower files processed
SME lendingConsumerCDFIsCommercial
Trusted by lenders across emerging markets
Amihan Credit
Pesolo
Naya Finance
TrustLine
Sari Capital
Habi Lending
Verda BNPL
Where credit teams get stuck today

Files sit incomplete. Good borrowers walk.

In emerging markets, the bottleneck isn't modeling — it's the file. Borrowers send messy documents, formats are local, and the bureau gives you nothing.

Credit officers spend most of their week reading, keying, and reconciling. The real decision takes minutes.

01

Files arrive incomplete

Borrowers send screenshots, photos, partial PDFs. Officers chase missing pages for days.

02

Documents don't fit your tools

GCash screenshots, KTP cards, CFDIs — none of them play nicely with form-based intake.

03

Reconciliation is manual

Cross-checking declared revenue against bank inflows happens in spreadsheets, by hand.

04

Verification is uneven

Tampering checks are inconsistent across analysts. Risk slips through the cracks.

Time to decision

From days to minutes, on the same file.

Same SME application, same files. LendTrace collapses every stage that doesn't require human judgment.

faster end to end
70%
lower cost per file
A single SME application
Manual LendTrace
Collect & complete the file3 days / live
Read & key in documents5 hrs / 1.4s ea
Reconcile & fraud-check4 hrs / auto
Draft the credit memo3 hrs / 30s
Officer makes the call — minutes, where it belongs
The platform

Three AI layers. One underwriting stack.

Each layer is useful on its own, and they compose into a single pipeline that turns raw borrower documents into a cited, decision-ready memo.

Layer 01

LendTrace Capture

Reads any borrower document, on day one.

Vision-language models extract structured fields from screenshots, PDFs, photos at angles, and local-format documents. No templates. No fine-tuning per format.

Formats read · last hour42 docs
bank_statementpayslipmobile_bankingktpcomprobante+18 more
avg_extract_time
1.4s
field_accuracy
99.2%
tamper_checks
per-row
output
clean JSON
MS
Maria Santos
via WhatsApp
Hi Maria — could you send the last 3 months of GCash statements? Just screenshots are fine.
Sige po, sandali lang.
Eto na po, salamat!
gcash-jan.jpg+2 more
Got them, thanks. One more — your latest payslip?
File completeness78%
Layer 02

AI Credit Officer

Completes the file by talking to the borrower.

A polite, multilingual agent that follows up on missing pages over WhatsApp, SMS, or email — until the file is decision-ready.

Layer 03

AI Underwriter

Drafts a cited credit memo.

A structured recommendation with every claim cited back to a source line. Your team approves, declines, or overrides — every time.

Draft recommendationApprove· Confidence 92%· Policy 11/12· Fraud low · 0.04

Maria operates a small sari-sari store with consistent inflows. Bank statements reconcile with declared revenue within 1%. One minor inconsistency flagged for review.

DSCR
1.84×
cited: bank_stmt_q3.pdf
Identity
Verified
cited: ktp_front.jpg

Decision rests with credit officer. Always.

LendTrace never approves loans. Your team makes every call. Always.
Two flagship products
End-to-end platform

LendTrace Underwriter

The whole stack. Borrower documents in, cited credit memo out. Built for lenders who want one vendor, one contract, one source of truth.

Pairs with

LendTrace Capture

Just the document API. Drop it into your existing workflow, get clean structured fields and fraud flags back.

Explore LendTrace Capture →
01

Push borrower documents

Any format. Screenshots, PDFs, photos. Local IDs and bank statements supported.

02

Get deterministic signals

Structured JSON with cited sources, completeness scores, and tampering flags.

03

Generate a cited memo

An auditable recommendation with every line traceable back to a source document.

One stack. Every scenario you underwrite.

01

Microfinance & small-ticket consumer

High volume, thin files, lots of mobile money. Capture reads it; the API gives a yes/no signal in under a second.

POST /v1/decisions · 920ms
payslip_aug.jpg ✓ EXTRACTED
bank_stmt_q3.pdf ⚑ TAMPERED
Built with: Capture · AI Underwriter
02

SME & commercial lending

Heavy documents, real reconciliation, real memos. Drafted in 30 seconds, your officer signs.

PDF
CREDIT_MEMO.pdf · drafted
Built with: Capture · AI Credit Officer · AI Underwriter
03

CDFI & community lending

Mission-driven institutions need consistent, auditable decisioning. Deploy in your VPC, on your terms.

SOC 2 Type IIIN PROGRESS
VPC deploymentAVAILABLE
Built with: Full stack
Decision-ready files

Decision-ready files, not raw extractions.

We don't hand you a JSON dump. We hand you a structured applicant profile, every claim cited back to a source line, ready for your officer to read and sign.

Your team spends time on judgment, not bookkeeping.

Maria Santos
SME loan · ₱500,000 · Quezon City, PH
DECISION-READY
Document review
  • Identity verified · ktp_front.jpg
  • Address confirmed · barangay_permit.pdf
  • 3 months bank statements reconciled
Financials & capacity
  • DSCR 1.84× · bank_stmt_q3.pdf
  • Monthly inflow ₱52,310 · gcash-jan.jpg + 2
  • Stable 9-month trend
For your review

One minor inconsistency: declared revenue ₱2.4M vs. AFS ₱2.38M. ~1% variance — likely rounding.

LendTrace's draft
Approve recommended
Open full memo →
Why now

AI-native underwriting in emerging markets just became possible.

For the first time, vision-language models can read messy local-format financial documents — GCash screenshots, KTP cards, CFDIs — fluently and reliably, in the wild.

The bottleneck of underwriting in emerging markets has always been the file. That bottleneck is now a software problem.

Built for the markets where lending is hardest.

We started where the documents are messiest and the bureaus thinnest. Three regions live, more on the way.

SEALIVE

Southeast Asia

  • Philippines · GCash · CIBI
  • Indonesia · KTP · OVO
Explore →
LATAMLIVE

Latin America

  • Mexico · CFDI · INE
Explore →
AfricaLIVE

Africa

  • South Africa · SARS · FICA
Explore →
Built by a team from
Stanford University
Apple
Microsoft
Stanford AI Lab
Tesla
NVIDIA
EFBacked by Entrepreneur First

Enterprise-ready by default. A platform that fits inside your credit policy.

i
Customer-specific deployments
Single-tenant or VPC, your choice.
ii
Provider allowlisting
Pin which model providers you allow.
iii
Lender-controlled decisioning
We never auto-approve. Officers decide.
iv
SOC 2 · ISO 27001
Type II in progress; ISO 27001 in audit.
v
Data stays in-region
PH · ID · MX · ZA residency available.