DebtLens

Every debt, every currency, one honest number

NEXT.JS
REACT 19
TYPESCRIPT
POSTGRES
DRIZZLE
ZITADEL

Summary

DebtLens pulls debts across institutions, currencies, and statement formats into one view — amortization, payoff scenarios, utilization, debt-to-income — computed deterministically to the cent. Next.js and Postgres, a pure-function math core, and statement parsing for PDF and CSV. In development.

A · Problem

Debt hides in fragments — a card here, a line of credit there, two currencies, a dozen statement layouts. No single view tells you what you actually owe, or the date you'd finally be free of it.

B · Approach

A side-effect-free calculation engine: money as integer cents, rates as integer basis points, so amortization schedules and payoff timelines come out deterministic and testable. Statements ingest through fuzzy CSV mapping and PDF parsing; Bank of Canada rates reconcile multi-currency balances; scheduled jobs handle reminders and refreshes.

C · Artifacts

React components: 159
Postgres tables (Drizzle): 13
math models, each unit-tested: 4
currencies reconciled — CAD · USD: 2

Amortization, payoff, utilization, and debt-to-income models each behind their own test suite; CSV and PDF statement parsers with a rule-based categorizer; S3 presigned uploads, Upstash rate limiting, Zitadel auth, and Vercel cron for daily FX and reminders.

D · Outcome

A precise, multi-currency picture of what you owe and when it ends — deterministic enough to act on. In development.

AlgoMotion