Skip to content
erzurumlu.dev
← All projects

FinTrack.

Your budget stays with you.

Personal budget tracking in one HTML file with zero dependencies.

Role
Design and development
Scope
Solo project
Year
2026
FinTrackApril 2026

Total balance

₺0

Income₺0
Expenses₺0

Savings goal

New computer

₺8.000 / ₺15.0000%

Category breakdown

Housing32%Food22%Transport14%Leisure11%Health9%Other12%

Recent transactions

Salary+28.500
Groceries−1.250
Electricity−780
Coffee−320
Illustrative interface · Sample data

No account. No budget in the cloud.

A lightweight budget tracker that runs locally in your browser. Your data stays on your device.

0lines
~0KB
0dependencies

How I built it.

A simple, solid structure running entirely in your browser.

One HTML file
Canvas 2D
localStorage

What it can do.

Income and expenses

Track transactions and recurring subscriptions.

Category analysis

See spending distribution in Canvas 2D charts.

Savings goals

Set goals in lira, dollar, euro and gram gold.

CSV export

Export your records whenever you need them.

Monthly spendingThis month
₺0Total spending
Housing0%
Food0%
Transport0%
Leisure0%
Health0%
Other0%
What I learned

“Understand what a library does before using it.”

Drawing charts directly with Canvas 2D helped me understand what a chart library does and what it hides.

Tech stack.

  • HTML
  • CSS
  • JavaScript
  • Canvas 2D
  • localStorage
Technical notes and project scope

The problem and my role

Most personal budgeting apps ask you either to create an account or to send your spending data to the cloud — yet a budget is among the most private data a person has. The question was: how small can a budgeting tool be if it sends data nowhere and needs no installation?

Entirely mine, from design to code. Written for a Financial Technologies course.

Implementation decisions

  • I wrote the whole app in a single HTML file: 639 lines, about 33 KB, no build step. You double-click it and it runs.
  • Data stays in the browser’s localStorage and never reaches a server. The one outbound request is for live exchange rates.
  • I drew the donut chart with Canvas 2D rather than Chart.js: slices computed as angles through ctx.arc(), then hollowed into a ring with a second circle.
  • For spending warnings I wrote a rule-based engine — rent-to-income ratio, savings percentage, concentration in discretionary categories.

Outcome and learning

Income and expense tracking, category analysis, recurring subscriptions, savings goals in four units (lira, dollar, euro, gram gold), CSV export and a dark theme — all with zero dependencies.

Installing Chart.js takes three minutes; drawing the ring yourself with trigonometry takes an afternoon. But afterwards you know what a charting library actually does, where it slows down and what it hides. Adding a dependency is sometimes outsourcing your own thinking.

The app’s "Financial Intelligence Assistant" is not an AI model but an engine of hand-written rules. The name is ambitious; what it does is threshold checking.