Credit scoring · KYC · KYB

Turn a borrower's documents into a lending decision.

Send an M-Pesa or bank statement, an ID and a selfie — whatever you have — and get one comparable apenia_score from 0–900, a risk band, an affordability figure and a recommended limit. Plus identity verification, liveness and credit-bureau checks, behind one API.

Data-agnostic · statements, collateral, KYCExplainable · every point traces to a reasonSandbox-first · try before you bill
One integration

Everything a lending decision needs

Each capability is metered independently — activate only what you use, priced per call.

Statement analysis

Score M-Pesa, bank and collateral documents into one 0–900 apenia_score with a full affordability breakdown.

Identity & liveness

Face-match a selfie to an ID with active liveness — smile, turn, nod — beating photo and screen spoofing.

Credit bureau

Pull a TransUnion or Metropol report, resolve alternative phones, and submit loans & repayments back to the bureau.

Voice biometrics

Enroll a borrower's voiceprint at onboarding and re-verify them by voice at loan time.

How it works

From a raw document to a decision, end to end

Your product stays in front of the borrower; Apenia sits behind it.

1

The documents are the input

Your customer has an M-Pesa or bank statement, maybe an ID photo and selfie. That's the raw material every decision starts from.

2

Your system makes the call

You integrate from your backend — your loan app, your form. A drop-in widget and hosted onboarding links exist if you'd rather not build the UI.

3

Authenticate & check funds

Every call carries your key + secret (or an OAuth token), applies your rate limit, and confirms your prepaid Float wallet covers premium calls.

4

Two brains, kept separate

A model extracts data from any document; a deterministic, auditable engine computes the score — income, affordability, debt, stability, behaviour.

5

Results come back four ways

In the JSON response, as a signed webhook to your callback, a debit to your Float wallet, and a logged transaction in the console — at once.

6

It sharpens with use

Report loan outcomes as they mature; once enough accumulate, the score's weights recalibrate against real defaults — staying transparent throughout.

The design principle

Two brains, kept separate — so the decision is never a black box.

Extraction is fuzzy, so a model does it. The decision is deterministic and explainable, so the engine does it — and every point traces to a named reason a regulator can audit. That separation is what makes Apenia fair-lending defensible, not just accurate.

189
M-Pesa only
550
+ verified salary
Fusing a verified bank salary re-rates a borrower from Catastrophic to Moderate — exactly the lift a lender wants.
Pricing

A prepaid Float wallet, priced per call

Top up by M-Pesa; each successful call debits its price and writes a line to your statement. High-volume? Take a monthly plan with an included quota and per-1,000 overage.

Statement analysis
Score a bank / M-Pesa statement
KES 100
ID + selfie match
Verify a borrower is who they say
KES 80
CRB check
Pull a credit-bureau report
KES 120
Billing is tied to your app, not your login method — it works the same however you authenticate.

Try it in sandbox — free — before you bill a shilling.

Read the API reference, mint a key, and score your first statement in minutes.