← bankstract

For lending fintechs

Bank statement parsing for underwriting

Turn a Nigerian bank statement PDF into clean transactions, balances, and account metadata over one API call. NDPR-aware redaction in the same call.

Built for a lending data pipeline

The ask is to trust an API with underwriting data. Here is why you can.

Open source engine

The parser is MIT-licensed. Read it, run it, audit the exact code that touches your data.

View the engine on GitHub

Self-hostable

Run the whole stack inside your own infrastructure. Nothing has to leave your network.

Self-host guide

Processed in memory

The statement is parsed in memory. It is never written to disk and never logged.

NDPR-aware redaction

Names, account numbers, and BVN masked in the same call. No second pipeline.

Metadata-only audit log

We record that a file parsed. Never what was in it. No transactions, no balances.

How it fits

  1. 01

    Applicant uploads a statement

    A PDF from any covered bank. The same file your underwriters read by hand today.

  2. 02

    One API call parses it

    Transactions, balances, and account metadata come back structured and typed. Every parse runs a balance check.

  3. 03

    You underwrite on clean data

    No fragile in-house parser to maintain. It handles the statements a direct-connect cannot reach, and the ones customers upload by hand.

curl -X POST https://bankstract.logickoder.dev/v1/parse \
  -H "Authorization: Bearer bsk_live_..." \
  -F "pdf=@statement.pdf"

Coverage

Live today: PalmPay, First Bank, Zenith, and Opay. New banks and format drift ship in about 48 hours. If a bank you need is not covered, that is exactly what a design partner sets the priority on.

Start on the free tier: 25 parses a month, no card. Paid tiers from ₦9,500/mo. See pricing.

Parse your first statement

Free demo, no signup. Self-host the whole stack, or use the cloud when paid tiers ship.

Need volume or NDPR audit support? Talk to sales.