I’m Tolga, a developer building web, Mac and Windows apps.
Useful software.
Clear code.
New possibilities with AI.
I work with JavaScript, TypeScript, React and Swift. Right now I’m building Ajandam, a free personal finance, calendar and notes app for Mac and Windows.
Ajandam
Personal finance, calendar and notes in one app, for Mac and Windows.
Free for macOS 12 or later and Windows 10 and 11.
On Mac, Ajandam is also on the Mac App Store.
In Turkey, Ajandam is also on the Microsoft Store.
Made for people in Turkey, so the interface is in Turkish. Accounts, credit cards, installments, budgets, net worth, an agenda and notes live in one window, and your records stay on your computer.
-
Your records stay on your computer
Accounts, credit cards with statement and due dates, installments, budgets and net worth. No account, no server.
-
AI without API keys
Apple Intelligence on the Mac, local models (Ollama, LM Studio), sign in with ChatGPT, or connect Claude over MCP. Nothing is saved without your approval.
-
Statement import
PDF, Excel and CSV bank statements, reconciled against the statement total, with AI-assisted reading and category suggestions.
-
Made for Turkey
A CPI-based inflation report, the rent increase cap, central bank card rates, public holidays and a tax calendar.
On the Mac, Ajandam is a native shell written in Swift (AppKit and WebKit); on Windows, it is built with Tauri (Rust and WebView2). Both run the same interface, written in plain JavaScript. Signed releases and one-click updates are built with GitHub Actions.
What I work on
Web, Mac and Windows apps
Practical interfaces with JavaScript, TypeScript and React, native Mac apps with Swift, and Windows apps with Tauri.
AI and developer tools
Exploring how AI can support prototyping, learning and development workflows.
Software quality
Readable code, thoughtful user experiences and useful documentation.
Tools I use most
- JavaScript
- TypeScript
- React
- Swift
- Rust
- Tauri
- macOS
- GitHub Actions
How I build
-
Understand the problem
Who is it for, and what do they actually need?
-
Build a focused version
Solve the core of it first and leave out the rest.
-
Test
Check it against real use, not only the happy path.
-
Improve
Refine what I learned, then start again from step one.
I value code I can understand, explain and maintain. My interest in AI is grounded in that same approach: explore ideas, question the output and verify the result.